#MCS00 @inproceedings{Benediktsson00Consensus, author = {Jon Benediktsson and Johannes R. Sveinsson}, title = {Consensus Based Classification of Multisource Remote Sensing Data.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {280-289}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570280.htm}, } @inproceedings{Bruzzone00Combining, author = {Lorenzo Bruzzone and Roberto Cossu and Diego Fern{\'a}ndez Prieto}, title = {Combining Parametric and Nonparametric Classifiers for an Unsupervised Updating of Land-Cover Maps.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {290-299}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570290.htm}, } @inproceedings{Cappelli00Combining, author = {Raffaele Cappelli and Dario Maio and Davide Maltoni}, title = {Combining Fingerprint Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {351-361}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570351.htm}, } @INPROCEEDINGS{Cohen00Hybrid, author = {Shimon Cohen and Nathan Intrator}, title = {A Hybrid Projection Based and Radial Basis Function Architecture.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {147-156}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570147.htm}, } @INPROCEEDINGS{Conversano00Supervised, author = {Claudio Conversano and Roberta Siciliano and Francesco Mola}, title = {Supervised Classifier Combination through Generalized Additive Multi-model.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {167-176}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570167.htm}, } @INPROCEEDINGS{Cordella00Cascaded, author = {Luigi P. Cordella and Pasquale Foggia and Carlo Sansone and Francesco Tortorella and Mario Vento}, title = {A Cascaded Multiple Expert System for Verification.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {330-339}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570330.htm}, } @INPROCEEDINGS{Dietterich00Ensemble, author = {Thomas G. Dietterich}, title = {Ensemble Methods in Machine Learning.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {1-15}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570001.htm}, } @INPROCEEDINGS{Diez00Applying, author = {Juan J. Rodr\'{\i}guez Diez and Carlos Alonso Gonz{\'a}lez}, title = {Applying Boosting to Similarity Literals for Time Series Classification.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {210-219}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570210.htm}, } @INPROCEEDINGS{Duin00Experiments, author = {Robert P. W. Duin and David M. J. Tax}, title = {Experiments with Classifier Combining Rules.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {16-29}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570016.htm}, } @INPROCEEDINGS{Fanelli00Modular, author = {Anna Maria Fanelli and Giovanna Castellano and C. Alessandro Buscicchio}, title = {A Modular Neuro-Fuzzy Network for Musical Instruments Classification.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {372-382}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570372.htm}, } @INPROCEEDINGS{Froeba00Statistical, author = {Bernhard Fr{\"o}ba and Constanze Rothe and Christian K{\"u}blbeck}, title = {Statistical Sensor Calibration for Fusion of Different Classifiers in a Biometric Person Recognition Framework.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {362-371}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570362.htm}, } @INPROCEEDINGS{Furlanello00Boosting, author = {Cesare Furlanello and Stefano Merler}, title = {Boosting of Tree-Based Classifiers for Predictive Risk Modeling in GIS.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {220-229}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570220.htm}, } @INPROCEEDINGS{Giacinto00Dynamic, author = {Giorgio Giacinto and Fabio Roli}, title = {Dynamic Classifier Selection.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {177-189}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570177.htm}, } @INPROCEEDINGS{Griffith00Self, author = {Niall Griffith and Derek Partridge}, title = {Self-Organizing Decomposition of Functions in the Context of a Unified Framework for Multiple Classifier Systems.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {250-259}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570250.htm}, } @INPROCEEDINGS{Grim00Combining, author = {Jiri Grim and Josef Kittler and Pavel Pudil and Petr Somol}, title = {Combining Multiple Classifiers in Probabilistic Neural Networks.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {157-166}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570157.htm}, } @INPROCEEDINGS{Happel00Analysis, author = {Mark D. Happel and Peter Bock}, title = {Analysis of a Fusion Method for Combining Marginal Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {137-146}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570137.htm}, } @INPROCEEDINGS{Ho00Complexity, author = {Tin Kam Ho}, title = {Complexity of Classification Problems and Comparative Advantages of Combined Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {97-106}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570097.htm}, } @INPROCEEDINGS{Ianakiev00Architecture, author = {Krassimir G. Ianakiev and Venu Govindaraju}, title = {Architecture for Classifier Combination Using Entropy Measures.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {340-350}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570340.htm}, } @INPROCEEDINGS{Impedovo00Evaluation, author = {Sebastiano Impedovo and A. Salzo}, title = {A New Evaluation Method for Expert Combination in Multi-expert System Designing.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {230-239}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570230.htm}, } @INPROCEEDINGS{Jiang00Some, author = {Wenxin Jiang}, title = {Some Results on Weakly Accurate Base Learners for Boosting Regression and Classification.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {87-96}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570087.htm}, } @INPROCEEDINGS{Jiang00Classifier, author = {Xiaoyi Jiang and Keren Yu and Horst Bunke}, title = {Classifier Combination for Grammar-Guided Sentence Recognition.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {383-392}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570383.htm}, } @PROCEEDINGS{Kittler00Multiple, editor = {Josef Kittler and Fabio Roli}, title = {Multiple Classifier Systems, First International Workshop, MCS 2000, Cagliari, Italy, June 21-23, 2000, Proceedings}, booktitle = {Multiple Classifier Systems}, publisher = {Springer}, series = {Lecture Notes in Computer Science}, volume = 1857, year = 2000, isbn = {3-540-67704-6}, } @INPROCEEDINGS{Kleinberg00Mathematically, author = {Eugene M. Kleinberg}, title = {A Mathematically Rigorous Foundation for Supervised Learning.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {67-76}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570067.htm}, } @INPROCEEDINGS{Kumar00Hierarchical, author = {Shailesh Kumar and Joydeep Ghosh and Melba M. Crawford}, title = {A Hierarchical Multiclassifier System for Hyperspectral Data Analysis.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {270-279}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570270.htm}, } @INPROCEEDINGS{Kumazawa00Shape, author = {Itsuo Kumazawa}, title = {Shape Matching and Extraction by an Array of Figure-and-Ground Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {393-402}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570393.htm}, } @INPROCEEDINGS{Lam00Classifier, author = {Louisa Lam}, title = {Classifier Combinations: Implementations and Theoretical Issues.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {77-86}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570077.htm}, } @INPROCEEDINGS{Latinne00Different, author = {Patrice Latinne and Olivier Debeir and Christine Decaestecker}, title = {Different Ways of Weakening Decision Trees and Their Impact on Classification Accuracy of DT Combination.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {200-209}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570200.htm}, } @INPROCEEDINGS{Lecce00Multi, author = {Vincenzo Di Lecce and Giovanni Dimauro and Andrea Guerriero and Sebastiano Impedovo and Giuseppe Pirlo and A. Salzo}, title = {A Multi-expert System for Dynamic Signature Verification.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {320-329}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570320.htm}, } @INPROCEEDINGS{Masulli00Effectiveness, author = {Francesco Masulli and Giorgio Valentini}, title = {Effectiveness of Error Correcting Output Codes in Multiclass Learning Problems.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {107-116}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570107.htm}, } @INPROCEEDINGS{Pekalska00Combining, author = {Elzbieta Pekalska and Marina Skurichina and Robert P. W. Duin}, title = {Combining Fisher Linear Discriminants for Dissimilarity Representations.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {117-126}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570117.htm}, } @INPROCEEDINGS{Sharkey00Test, author = {Amanda J. C. Sharkey and Noel E. Sharkey and Uwe Gerecke and Gopinath Odayammadath Chandroth}, title = {The "Test and Select" Approach to Ensemble Combination.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {30-44}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570030.htm}, } @INPROCEEDINGS{Skurichina00Boosting, author = {Marina Skurichina and Robert P. W. Duin}, title = {Boosting in Linear Discriminant Analysis.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {190-199}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570190.htm}, } @INPROCEEDINGS{Slavik00Lexicon, author = {Petr Slav\'{\i}k and Venu Govindaraju}, title = {Use of Lexicon Density in Evaluating Word Recognizers.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {310-319}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570310.htm}, } @INPROCEEDINGS{Srihari00Survey, author = {Sargur N. Srihari}, title = {A Survey of Sequential Combination of Word Recognizers in Handwritten Phrase Recognition at CEDAR.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {45-51}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570045.htm}, } @INPROCEEDINGS{Suen00Multiple, author = {Ching Y. Suen and Louisa Lam}, title = {Multiple Classifier Combination Methodologies for Different Output Levels.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {52-66}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570052.htm}, } @INPROCEEDINGS{Takahashi00Learning, author = {Katsuhiko Takahashi and Atsushi Sato}, title = {A Learning Method of Feature Selection for Rough Classification.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {127-136}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570127.htm}, } @INPROCEEDINGS{Wan00Multiple, author = {Weijian Wan and Donald Fraser}, title = {A Multiple Self-Organizing Map Scheme for Remote Sensing Classification.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {300-309}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570300.htm}, } @INPROCEEDINGS{Wang00Diversity, author = {Wenjia Wang and Phillis Jones and Derek Partridge}, title = {Diversity between Neural Networks and Decision Trees for Building Multiple Classifier Systems.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {240-249}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570240.htm}, } @INPROCEEDINGS{Windeatt00Classifier, author = {Terry Windeatt}, title = {Classifier Instability and Partitioning.}, booktitle = {Multiple Classifier Systems}, year = 2000, pages = {260-269}, ee = {http://link.springer.de/link/service/series/0558/bibs/1857/18570260.htm}, } #MCS01 @inproceedings{Alkoot01Improving, author = {Fuad M. Alkoot and Josef Kittler}, title = {Improving Product by Moderating k-NN Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {429-439}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960429.htm}, } @INPROCEEDINGS{Briem01Boosting, author = {Gunnar Jakob Briem and Jon Atli Benediktsson and Johannes R. Sveinsson}, title = {Boosting, Bagging, and Consensus Based Classification of Multisource Remote Sensing Data.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {279-288}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960279.htm}, } @INPROCEEDINGS{Bruzzone01Robust, author = {Lorenzo Bruzzone and Roberto Cossu}, title = {A Robust Multiple Classifier System for a Partially Unsupervised Updating of Land-Cover Maps.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {259-268}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960259.htm}, } @INPROCEEDINGS{Chen01Averaging, author = {Dechang Chen and Jian Liu}, title = {Averaging Weak Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {119-125}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960119.htm}, } @INPROCEEDINGS{Cohen01Automatic, author = {Shimon Cohen and Nathan Intrator}, title = {Automatic Model Selection in a Hybrid Perceptron/Radial Network.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {440-454}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960440.htm}, } @INPROCEEDINGS{Dahmen01Combined, author = {J{\"o}rg Dahmen and Daniel Keysers and Hermann Ney}, title = {Combined Classification of Handwritten Digits Using the 'Virtual Test Sample Method'.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {109-118}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960109.htm}, } @INPROCEEDINGS{Dietrich01Classification, author = {Christian Dietrich and Friedhelm Schwenker and G{\"u}nther Palm}, title = {Classification of Time Series Utilizing Temporal and Decision Fusion.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {378-387}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960378.htm}, } @INPROCEEDINGS{Diez01Learning, author = {Juan J. Rodr\'{\i}guez Diez and Carlos Alonso Gonz{\'a}lez}, title = {Learning Classification RBF Networks by Boosting.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {43-52}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960043.htm}, } @INPROCEEDINGS{Dolenko01Solar, author = {S. A. Dolenko and Yu. V. Orlov and I. G. Persiantsev and Ju. S. Shugai and A. V. Dmitriev and A. V. Suvorova and I. S. Veselovsky}, title = {Solar Wind Data Analysis Using Self-Organizing Hierarchical Neural Network Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {289-298}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960289.htm}, } @INPROCEEDINGS{Foggia01Automatic, author = {Pasquale Foggia and Carlo Sansone and Francesco Tortorella and Mario Vento}, title = {Automatic Classification of Clustered Microcalcifications by a Multiple Classifier System.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {208-217}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960208.htm}, } @INPROCEEDINGS{Fred01Finding, author = {Ana L. N. Fred}, title = {Finding Consistent Clusters in Data Partitions.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {309-318}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960309.htm}, } @INPROCEEDINGS{Froeba01Combination, author = {Bernhard Fr{\"o}ba and Walter Zink}, title = {On the Combination of Different Template Matching Strategies for Fast Face Detection.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {418-428}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960418.htm}, } @INPROCEEDINGS{Frossyniotis01Multi, author = {Dimitrios S. Frossyniotis and Andreas Stafylopatis}, title = {A Multi-SVM Classification System.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {198-207}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960198.htm}, } @INPROCEEDINGS{Fumera01Error, author = {Giorgio Fumera and Fabio Roli}, title = {Error Rejection in Linearly Combined Multiple Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {329-338}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960329.htm}, } @INPROCEEDINGS{Ghaderi01Least, author = {Reza Ghaderi and Terry Windeatt}, title = {Least Squares and Estimation Measures via Error Correcting Output Code.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {148-157}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960148.htm}, } @INPROCEEDINGS{Gini01Mixing, author = {Giuseppina C. Gini and Marco Lorenzini and Emilio Benfenati and Raffaella Brambilla and Luca Malv{\'e}}, title = {Mixing a Symbolic and a Subsymbolic Expert to Improve Carcinogenicity Prediction of Aromatic Compounds.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {126-135}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960126.htm}, } @INPROCEEDINGS{Grim01Information, author = {Jiri Grim and Josef Kittler and Pavel Pudil and Petr Somol}, title = {Information Analysis of Multiple Classifier Fusion.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {168-177}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960168.htm}, } @INPROCEEDINGS{Hand01Multiple, author = {David J. Hand and Niall M. Adams and Mark G. Kelly}, title = {Multiple Classifier Systems Based on Interpretable Linear Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {136-147}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960136.htm}, } @INPROCEEDINGS{Hartono01Learning, author = {Pitoyo Hartono and Shuji Hashimoto}, title = {Learning-Data Selection Mechanism through Neural Networks Ensemble.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {188-197}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960188.htm}, } @INPROCEEDINGS{Higgins01Application, author = {Jonathan E. Higgins and Tony J. Dodd and Robert I. Damper}, title = {Application of Multiple Classifier Techniques to Subband Speaker Identification with an HMM/ANN System.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {369-377}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960369.htm}, } @INPROCEEDINGS{Ho01Data, author = {Tin Kam Ho}, title = {Data Complexity Analysis for Classifier Combination.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {53-67}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960053.htm}, } @INPROCEEDINGS{Jorgensen01Feature, author = {Thomas Martini J{\o}rgensen and Christian Linneberg}, title = {Feature Weighted Ensemble Classifiers - A Modified Decision Scheme.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {218-227}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960218.htm}, } @INPROCEEDINGS{Kittler01Relationship, author = {Josef Kittler and Fuad M. Alkoot}, title = {Relationship of Sum and Vote Fusion Strategies.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {339-348}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960339.htm}, } @PROCEEDINGS{Kittler01Multiple, editor = {Josef Kittler and Fabio Roli}, title = {Multiple Classifier Systems, Second International Workshop, MCS 2001 Cambridge, UK, July 2-4, 2001, Proceedings}, booktitle = {Multiple Classifier Systems}, publisher = {Springer}, series = {Lecture Notes in Computer Science}, volume = 2096, year = 2001, isbn = {3-540-42284-6}, } @INPROCEEDINGS{Kuncheva01Complexity, author = {Ludmila I. Kuncheva and Fabio Roli and Gian Luca Marcialis and Catherine A. Shipp}, title = {Complexity of Data Subsets Generated by the Random Subspace Method: An Experimental Investigation.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {349-358}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960349.htm}, } @INPROCEEDINGS{Kuncheva01Feature, author = {Ludmila I. Kuncheva and Christopher J. Whitaker}, title = {Feature Subsets for Classifier Combination: An Enumerative Experiment.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {228-237}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960228.htm}, } @INPROCEEDINGS{Langdon01Genetic, author = {William B. Langdon and Bernard F. Buxton}, title = {Genetic Programming for Improved Receiver Operating Characteristics.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {68-77}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960068.htm}, } @INPROCEEDINGS{Latinne01Limiting, author = {Patrice Latinne and Olivier Debeir and Christine Decaestecker}, title = {Limiting the Number of Trees in Random Forests.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {178-187}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960178.htm}, } @INPROCEEDINGS{Luttrell01Self, author = {Stephen P. Luttrell}, title = {A Self-Organising Approach to Multiple Classifier Fusion.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {319-328}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960319.htm}, } @INPROCEEDINGS{Marti01Positional, author = {Urs-Viktor Marti and Horst Bunke}, title = {Use of Positional Information in Sequence Alignment for Multiple Classifier Combination.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {388-398}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960388.htm}, } @INPROCEEDINGS{Masulli01Dependence, author = {Francesco Masulli and Giorgio Valentini}, title = {Dependence among Codeword Bits Errors in ECOC Learning Machines: An Experimental Analysis.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {158-167}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960158.htm}, } @INPROCEEDINGS{Merler01Tuning, author = {Stefano Merler and Cesare Furlanello and Barbara Larcher and Andrea Sboner}, title = {Tuning Cost-Sensitive Boosting and Its Application to Melanoma Diagnosis.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {32-42}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960032.htm}, } @INPROCEEDINGS{Oza01Input, author = {Nikunj C. Oza and Kagan Tumer}, title = {Input Decimation Ensembles: Decorrelation through Dimensionality Reduction.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {238-247}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960238.htm}, } @INPROCEEDINGS{Pekalska01Combining, author = {Elzbieta Pekalska and Robert P. W. Duin}, title = {On Combining Dissimilarity Representations.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {359-368}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960359.htm}, } @INPROCEEDINGS{Prabhakar01Decision, author = {Salil Prabhakar and Anil K. Jain}, title = {Decision-Level Fusion in Fingerprint Verification.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {88-98}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960088.htm}, } @INPROCEEDINGS{Roli01Methods, author = {Fabio Roli and Giorgio Giacinto and Gianni Vernazza}, title = {Methods for Designing Multiple Classifier Systems.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {78-87}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960078.htm}, } @INPROCEEDINGS{Ruta01Application, author = {Dymitr Ruta and Bogdan Gabrys}, title = {Application of the Evolutionary Algorithms for Classifier Selection in Multiple Classifier Systems with Majority Voting.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {399-408}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960399.htm}, } @INPROCEEDINGS{Schwenker01Tree, author = {Friedhelm Schwenker and G{\"u}nther Palm}, title = {Tree-Structured Support Vector Machines for Multi-class Pattern Recognition.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {409-417}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960409.htm}, } @INPROCEEDINGS{Sirlantzis01Genetic, author = {Konstantinos Sirlantzis and Michael C. Fairhurst and Sanaul Hoque}, title = {Genetic Algorithms for Multi-classifier System Configuration: A Case Study in Character Recognition.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {99-108}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960099.htm}, } @INPROCEEDINGS{Skurichina01Bagging, author = {Marina Skurichina and Robert P. W. Duin}, title = {Bagging and the Random Subspace Method for Redundant Feature Spaces.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {1-10}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960001.htm}, } @INPROCEEDINGS{Smits01Combining, author = {Paul C. Smits}, title = {Combining Supervised Remote Sensing Image Classifiers Based on Individual Class Performances.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {269-278}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960269.htm}, } @INPROCEEDINGS{Tapia01Generalized, author = {Elizabeth Tapia and Jos{\'e} Carlos Gonz{\'a}lez and Julio Villena}, title = {A Generalized Class of Boosting Algorithms Based on Recursive Decoding Models.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {22-31}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960022.htm}, } @INPROCEEDINGS{Tax01Combining, author = {David M. J. Tax and Robert P. W. Duin}, title = {Combining One-Class Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {299-308}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960299.htm}, } @INPROCEEDINGS{Wickramaratna01Performance, author = {Jeevani Wickramaratna and Sean B. Holden and Bernard F. Buxton}, title = {Performance Degradation in Boosting.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {11-21}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960011.htm}, } @INPROCEEDINGS{Windridge01Classifier, author = {David Windridge and Josef Kittler}, title = {Classifier Combination as a Tomographic Process.}, booktitle = {Multiple Classifier Systems}, year = 2001, pages = {248-258}, ee = {http://link.springer.de/link/service/series/0558/bibs/2096/20960248.htm}, } #MCS02 @INPROCEEDINGS{Altincay02Post, author = {Hakan Altin\c{c}ay and M{\"u}beccel Demirekler}, title = {Post-processing of Classifier Outputs in Multiple Classifier Systems.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {159-168}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640159.htm}, } @INPROCEEDINGS{Benfenati02Combining, author = {Emilio Benfenati and Paolo Mazzatorta and Daniel Neagu and Giuseppina C. Gini}, title = {Combining Classifiers of Pesticides Toxicity through a Neuro-fuzzy Approach.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {293-303}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640293.htm}, } @INPROCEEDINGS{Caprile02Highlighting, author = {Bruno Caprile and Cesare Furlanello and Stefano Merler}, title = {Highlighting Hard Patterns via AdaBoost Weights Evolution.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {72-80}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640072.htm}, } @INPROCEEDINGS{Chawla02Distributed, author = {Nitesh V. Chawla and Lawrence O. Hall and Kevin W. Bowyer and Thomas E. Moore and W. Philip Kegelmeyer}, title = {Distributed Pasting of Small Votes.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {52-61}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640052.htm}, } @INPROCEEDINGS{Cohen02Forward, author = {Shimon Cohen and Nathan Intrator}, title = {Forward and Backward Selection in Regression Hybrid Network.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {98-107}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640098.htm}, } @INPROCEEDINGS{Cordella02Multi, author = {Luigi P. Cordella and Massimo De Santo and Gennaro Percannella and Carlo Sansone and Mario Vento}, title = {A Multi-expert System for Movie Segmentation.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {304-313}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640304.htm}, } @INPROCEEDINGS{Dzeroski02Stacking, author = {Saso Dzeroski and Bernard Zenko}, title = {Stacking with Multi-response Model Trees.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {201-211}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640201.htm}, } @INPROCEEDINGS{Ghosh02Multiclassifier, author = {Joydeep Ghosh}, title = {Multiclassifier Systems: Back to the Future.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {1-15}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640001.htm}, } @INPROCEEDINGS{Guenter02Generating, author = {Simon G{\"u}nter and Horst Bunke}, title = {Generating Classifier Ensembles from Multiple Prototypes and Its Application to Handwriting Recognition.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {179-188}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640179.htm}, } @INPROCEEDINGS{Janeliunas02Reduction, author = {Arunas Janeliunas and Sarunas Raudys}, title = {Reduction of the Boasting Bias of Linear Experts.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {242-251}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640242.htm}, } @INPROCEEDINGS{Kittler02Decision, author = {Josef Kittler and Marco Ballette and Jacek Czyz and Fabio Roli and Luc Vandendorpe}, title = {Decision Level Fusion of Intramodal Personal Identity Verification Experts.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {314-324}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640314.htm}, } @INPROCEEDINGS{Kuncheva02Using, author = {Ludmila I. Kuncheva and Christopher J. Whitaker}, title = {Using Diversity with Three Variants of Boosting: Aggressive, Conservative, and Inverse.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {81-90}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640081.htm}, } @INPROCEEDINGS{Lai02Combining, author = {Carmen Lai and David M. J. Tax and Robert P. W. Duin and Elzbieta Pekalska and Pavel Pacl\'{\i}k}, title = {On Combining One-Class Classifiers for Image Database Retrieval.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {212-221}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640212.htm}, } @INPROCEEDINGS{Masulli02Boosting, author = {Francesco Masulli and Matteo Pardo and Giorgio Sberveglieri and Giorgio Valentini}, title = {Boosting and Classification of Electronic Nose Data.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {262-271}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640262.htm}, } @INPROCEEDINGS{Minguillon02Classifier, author = {Juli{\`a} Minguill{\'o}n and Anne Rosemary Tate and Carles Ar{\'u}s and John R. Griffiths}, title = {Classifier Combination for In Vivo Magnetic Resonance Spectra of Brain Tumours.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {282-292}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640282.htm}, } @INPROCEEDINGS{Mola02Discriminant, author = {Francesco Mola and Roberta Siciliano}, title = {Discriminant Analysis and Factorial Multiple Splits in Recursive Partitioning for Data Mining.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {118-126}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640118.htm}, } @INPROCEEDINGS{Morgan02Adaptive, author = {Joseph T. Morgan and Alex Henneguelle and Melba M. Crawford and Joydeep Ghosh and Amy Neuenschwander}, title = {Adaptive Feature Spaces for Land Cover Classification with Limited Ground Truth Data.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {189-200}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640189.htm}, } @INPROCEEDINGS{Pekalska02Discussion, author = {Elzbieta Pekalska and Robert P. W. Duin and Marina Skurichina}, title = {A Discussion on the Classifier Projection Space for Classifier Combining.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {137-148}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640137.htm}, } @INPROCEEDINGS{Raudys02Multiple, author = {Sarunas Raudys}, title = {Multiple Classification Systems in the Context of Feature Extraction and Selection.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {27-41}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640027.htm}, } @INPROCEEDINGS{Roli02Analysis, author = {Fabio Roli and Giorgio Fumera}, title = {Analysis of Linear and Order Statistics Combiners for Fusion of Imbalanced Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {252-261}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640252.htm}, } @PROCEEDINGS{Roli02Multiple, editor = {Fabio Roli and Josef Kittler}, title = {Multiple Classifier Systems, Third International Workshop, MCS 2002, Cagliari, Italy, June 24-26, 2002, Proceedings}, booktitle = {Multiple Classifier Systems}, publisher = {Springer}, series = {Lecture Notes in Computer Science}, volume = 2364, year = 2002, isbn = {3-540-43818-1}, } @INPROCEEDINGS{Roli02Multimodal, author = {Fabio Roli and Josef Kittler and Giorgio Fumera and Daniele Muntoni}, title = {An Experimental Comparison of Classifier Fusion Rules for Multimodal Personal Identity Verification Systems.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {325-336}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640325.htm}, } @INPROCEEDINGS{Roli02Crisp, author = {Fabio Roli and Sarunas Raudys and Gian Luca Marcialis}, title = {An Experimental Comparison of Fixed and Trained Fusion Rules for Crisp Classifier Outputs.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {232-241}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640232.htm}, } @INPROCEEDINGS{Ruta02Measure, author = {Dymitr Ruta and Bogdan Gabrys}, title = {New Measure of Classifier Dependency in Multiple Classifier Systems.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {127-136}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640127.htm}, } @INPROCEEDINGS{Schettini02Content, author = {Raimondo Schettini and Carla Brambilla and Claudio Cusano}, title = {Content-Based Classification of Digital Photos.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {272-281}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640272.htm}, } @INPROCEEDINGS{Sharkey02Types, author = {Amanda J. C. Sharkey}, title = {Types of Multinet System.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {108-117}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640108.htm}, } @INPROCEEDINGS{Sirlantzis02Trainable, author = {Konstantinos Sirlantzis and Sanaul Hoque and Michael C. Fairhurst}, title = {Trainable Multiple Classifier Schemes for Handwritten Character Recognition.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {169-178}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640169.htm}, } @INPROCEEDINGS{Skurichina02Bagging, author = {Marina Skurichina and Ludmila Kuncheva and Robert P. W. Duin}, title = {Bagging and Boosting for the Nearest Mean Classifier: Effects of Sample Size on Diversity and Accuracy.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {62-71}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640062.htm}, } @INPROCEEDINGS{Valentini02Bias, author = {Giorgio Valentini and Thomas G. Dietterich}, title = {Bias-Variance Analysis and Ensembles of SVM.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {222-231}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640222.htm}, } @INPROCEEDINGS{Windeatt02Boosted, author = {Terry Windeatt and Gholamreza Ardeshir}, title = {Boosted Tree Ensembles for Solving Multiclass Problems.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {42-51}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640042.htm}, } @INPROCEEDINGS{Windridge02General, author = {David Windridge and Josef Kittler}, title = {On the General Application of the Tomographic Classifier Fusion Methodology.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {149-158}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640149.htm}, } @INPROCEEDINGS{Yang02Multistage, author = {Shuang Yang and Antony Browne and Phil D. Picton}, title = {Multistage Neural Network Ensembles.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {91-97}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640091.htm}, } @INPROCEEDINGS{Zhu02Support, author = {Ji Zhu and Trevor Hastie}, title = {Support Vector Machines, Kernel Logistic Regression and Boosting.}, booktitle = {Multiple Classifier Systems}, year = 2002, pages = {16-26}, ee = {http://link.springer.de/link/service/series/0558/bibs/2364/23640016.htm}, } #MCS03 @INPROCEEDINGS{AhmadCVS03, author = {Khurshid Ahmad and Matthew Casey and Bogdan Vrusias and Panagiotis Saragiotis}, title = {Combining Multiple Modes of Information Using Unsupervised Neural Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {236-245}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090236.htm}, } @INPROCEEDINGS{Aksela03, author = {Matti Aksela}, title = {Comparison of Classifier Selection Methods for Improving Committee Performance.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {84-93}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090084.htm}, } @INPROCEEDINGS{AlamRT03, author = {Hassan Alam and Ahmad Fuad Rezaur Rahman and Yuliya Tarnikova}, title = {Solving Problems Two at a Time: Classification of Web Pages Using a Generic Pair-Wise Multiple Classifier System.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {385-394}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090385.htm}, } @INPROCEEDINGS{Arenas-GarciaFS03, author = {J. Arenas-Garc\'{\i}a and An\'{\i}bal R. Figueiras-Vidal and Amanda J. C. Sharkey}, title = {The Beneficial Effects of Using Multi-net Systems That Focus on Hard Patterns.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {45-54}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090045.htm}, } @INPROCEEDINGS{AsdornwisedJ03, author = {Widhyakorn Asdornwised and Somchai Jitapunkul}, title = {Automatic Target Recognition Using Multiple Description Coding Models for Multiple Classifier Systems.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {336-345}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090336.htm}, } @INPROCEEDINGS{AyadK03, author = {Hanan Ayad and Mohamed S. Kamel}, title = {Finding Natural Clusters Using Multi-clusterer Combiner Based on Shared Nearest Neighbors.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {166-175}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090166.htm}, } @INPROCEEDINGS{BanfieldHBK03, author = {Robert E. Banfield and Lawrence O. Hall and Kevin W. Bowyer and W. Philip Kegelmeyer}, title = {A New Ensemble Diversity Measure Applied to Thinning Ensembles.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {306-316}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090306.htm}, } @INPROCEEDINGS{BaykutE03, author = {Alper Baykut and Ayt{\"u}l Er\c{c}il}, title = {Towards Automated Classifier Combination for Pattern Recognition.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {94-105}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090094.htm}, } @INPROCEEDINGS{BrownW03, author = {Gavin Brown and Jeremy L. Wyatt}, title = {Negative Correlation Learning and the Ambiguity Family of Ensemble Methods.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {266-275}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090266.htm}, } @INPROCEEDINGS{Christensen03, author = {Stefan W. Christensen}, title = {Ensemble Construction via Designed Output Distortion.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {286-295}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090286.htm}, } @INPROCEEDINGS{CohenI03, author = {Shimon Cohen and Nathan Intrator}, title = {A Study of Ensemble of Hybrid Networks with Strong Regularization.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {227-235}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090227.htm}, } @INPROCEEDINGS{Cutzu03, author = {Florin Cutzu}, title = {Polychotomous Classification with Pairwise Classifiers: A New Voting Principle.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {115-124}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090115.htm}, } @INPROCEEDINGS{DuanKCSP03, author = {Kaibo Duan and S. Sathiya Keerthi and Wei Chu and Shirish Krishnaj Shevade and Aun Neow Poo}, title = {Multi-category Classification by Soft-Max Combination of Binary Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {125-134}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090125.htm}, } @INPROCEEDINGS{EstruchFHR03, author = {Vicent Estruch and C{\'e}sar Ferri and Jos{\'e} Hern{\'a}ndez-Orallo and M. Jos{\'e} Ram\'{\i}rez-Quintana}, title = {Beam Search Extraction and Forgetting Strategies on Shared Ensembles.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {206-216}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090206.htm}, } @INPROCEEDINGS{FumeraR03, author = {Giorgio Fumera and Fabio Roli}, title = {Linear Combiners for Classifier Fusion: Some Theoretical and Experimental Results.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {74-83}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090074.htm}, } @INPROCEEDINGS{GiacintoRD03, author = {Giorgio Giacinto and Fabio Roli and Luca Didaci}, title = {A Modular Multiple Classifier System for the Detection of Intrusions in Computer Networks.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {346-355}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090346.htm}, } @INPROCEEDINGS{GunterB03, author = {Simon G{\"u}nter and Horst Bunke}, title = {New Boosting Algorithms for Classification Problems with Large Number of Classes Applied to a Handwritten Word Recognition Task.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {326-335}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090326.htm}, } @INPROCEEDINGS{InoueN03, author = {Hirotaka Inoue and Hiroyuki Narihisa}, title = {Improving Performance of a Multiple Classifier System Using Self-generating Neural Networks.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {256-265}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090256.htm}, } @INPROCEEDINGS{JaserKC03, author = {Edward Jaser and Josef Kittler and William J. Christmas}, title = {Building Classifier Ensembles for Automatic Sports Classification.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {366-374}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090366.htm}, } @INPROCEEDINGS{KamelW03, author = {Mohamed S. Kamel and Nayer M. Wanas}, title = {Data Dependence in Combining Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {1-14}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090001.htm}, } @INPROCEEDINGS{KittlerAW03, author = {Josef Kittler and Alireza Ahmadyfard and David Windridge}, title = {Serial Multiple Classifier Systems Exploiting a Coarse to Fine Output Coding.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {106-114}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090106.htm}, } @INPROCEEDINGS{KoB03, author = {Jaepil Ko and Hyeran Byun}, title = {Binary Classifier Fusion Based on the Basic Decomposition Methods.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {146-155}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090146.htm}, } @INPROCEEDINGS{Kuncheva03, author = {Ludmila I. Kuncheva}, title = {Error Bounds for Aggressive and Conservative AdaBoost.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {25-34}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090025.htm}, } @INPROCEEDINGS{LewittP03, author = {Michael Lewitt and Robi Polikar}, title = {An Ensemble Approach for Data Fusion with Learn++.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {176-185}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090176.htm}, } @INPROCEEDINGS{Luttrell03, author = {Stephen P. Luttrell}, title = {A Markov Chain Approach to Multiple Classifier Fusion.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {217-226}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090217.htm}, } @INPROCEEDINGS{Magee03, author = {Derek R. Magee}, title = {A Sequential Scheduling Approach to Combining Multiple Object Classifiers Using Cross-Entropy.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {135-145}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090135.htm}, } @INPROCEEDINGS{McDonaldHE03, author = {Ross A. McDonald and David J. Hand and Idris A. Eckley}, title = {An Empirical Comparison of Three Boosting Algorithms on Real Data Sets with Artificial Class Noise.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {35-44}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090035.htm}, } @INPROCEEDINGS{Oza03, author = {Nikunj C. Oza}, title = {Boosting with Averaged Weight Vectors.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {15-24}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090015.htm}, } @INPROCEEDINGS{OzaTTH03, author = {Nikunj C. Oza and Kagan Tumer and Irem Y. Tumer and Edward M. Huff}, title = {Classification of Aircraft Maneuvers for Fault Detection.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {375-384}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090375.htm}, } @INPROCEEDINGS{RaudysR03, author = {Sarunas Raudys and Fabio Roli}, title = {The Behavior Knowledge Space Fusion Method: Analysis of Generalization Error and Strategies for Performance Improvement.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {55-64}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090055.htm}, } @INPROCEEDINGS{RaudysSB03, author = {Sarunas Raudys and Ray L. Somorjai and Richard Baumgartner}, title = {Reducing the Overconfidence of Base Classifiers when Combining Their Decisions.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {65-73}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090065.htm}, } @INPROCEEDINGS{SantosVAF03, author = {Rafael Valle dos Santos and Marley B. R. Vellasco and Fredy Artola and S{\'e}rgio da Fontoura}, title = {Neural Net Ensembles for Lithology Recognition.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {246-255}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090246.htm}, } @INPROCEEDINGS{SirlantzisHF03, author = {Konstantinos Sirlantzis and Sanaul Hoque and Michael C. Fairhurst}, title = {Input Space Transformations for Multi-classifier Systems Based on n-tuple Classifiers with Application to Handwriting Recognition.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {356-365}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090356.htm}, } @INPROCEEDINGS{TapiaGG03, author = {Elizabeth Tapia and Jos{\'e} Carlos Gonz{\'a}lez and L. Javier Garc\'{\i}a-Villalba}, title = {Good Error Correcting Output Codes for Adaptive Multiclass Learning.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {156-165}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090156.htm}, } @INPROCEEDINGS{VelekJN03, author = {Ondrej Velek and Stefan J{\"a}ger and Masaki Nakagawa}, title = {Accumulated-Recognition-Rate Normalization for Combining Multiple On/Off-Line Japanese Character Classifiers Tested on a Large Database.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {196-205}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090196.htm}, } @INPROCEEDINGS{VerbaetenA03, author = {Sofie Verbaeten and Anneleen Van Assche}, title = {Ensemble Methods for Noise Elimination in Classification Problems.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {317-325}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090317.htm}, } @INPROCEEDINGS{WilczokL03, author = {Elke Wilczok and Wolfgang Lellmann}, title = {Design and Evaluation of an Adaptive Combination Framework for OCR Result Strings.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {395-404}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090395.htm}, } @INPROCEEDINGS{WindeattGA03, author = {Terry Windeatt and Reza Ghaderi and Gholamreza Ardeshir}, title = {Spectral Coefficients and Classifier Correlation.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {276-285}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090276.htm}, } @PROCEEDINGS{mcs2003, editor = {Terry Windeatt and Fabio Roli}, title = {Multiple Classifier Systems, 4th International Workshop, MCS 2003, Guilford, UK, June 11-13, 2003, Proceedings}, booktitle = {Multiple Classifier Systems}, publisher = {Springer}, series = {Lecture Notes in Computer Science}, volume = 2709, year = 2003, isbn = {3-540-40369-8}, } @INPROCEEDINGS{WindridgeK03, author = {David Windridge and Josef Kittler}, title = {The Practical Performance Characteristics of Tomographically Filtered Multiple Classifier Fusion.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {186-195}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090186.htm}, } @INPROCEEDINGS{ZouariHLA03, author = {H{\'e}la Zouari and Laurent Heutte and Yves Lecourtier and Adel M. Alimi}, title = {Simulating Classifier Outputs for Evaluating Parallel Combination Methods.}, booktitle = {Multiple Classifier Systems}, year = 2003, pages = {296-305}, ee = {http://link.springer.de/link/service/series/0558/bibs/2709/27090296.htm}, } #MCS04 @INPROCEEDINGS{AyadBK04, author = {Hanan Ayad and Otman A. Basir and Mohamed Kamel}, title = {A Probabilistic Model Using Information Theoretic Measures for Cluster Ensembles.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {144-153}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=144}, } @INPROCEEDINGS{BanfieldHBBKE04, author = {Robert E. Banfield and Lawrence O. Hall and Kevin W. Bowyer and Divya Bhadoria and W. Philip Kegelmeyer and Steven Eschrich}, title = {A Comparison of Ensemble Creation Techniques.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {223-232}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=223}, } @INPROCEEDINGS{BonissoneGY04, author = {Piero P. Bonissone and Kai Goebel and Weizhong Yan}, title = {Classifier Fusion Using Triangular Norms.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {154-163}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=154}, } @INPROCEEDINGS{CaprileMFJ04, author = {Bruno Caprile and Stefano Merler and Cesare Furlanello and Giuseppe Jurman}, title = {Exact Bagging with k-Nearest Neighbour Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {72-81}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=72}, } @INPROCEEDINGS{ChenKJ04, author = {Lei Chen and Mohamed Kamel and Ju Jiang}, title = {A Modular System for the Classification of Time Series Data.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {134-143}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=134}, } @INPROCEEDINGS{CordellaLS04, author = {Luigi P. Cordella and Alessandro Limongiello and Carlo Sansone}, title = {Network Intrusion Detection by a Multi-stage Classification System.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {324-333}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=324}, } @INPROCEEDINGS{DaraK04, author = {Rozita A. Dara and Mohamed S. Kamel}, title = {Sharing Training Patterns among Multiple Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {243-252}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=243}, } @INPROCEEDINGS{DidaciG04, author = {Luca Didaci and Giorgio Giacinto}, title = {Dynamic Classifier Selection by Adaptive k-Nearest-Neighbourhood Rule.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {174-183}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=174}, } @INPROCEEDINGS{DiegoMM04, author = {Isaac Mart\'{\i}n de Diego and Javier M. Moguerza and Alberto Mu{\~n}oz}, title = {Combining Kernel Information for Support Vector Classification.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {102-111}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=102}, } @INPROCEEDINGS{EstruchFHR04, author = {Vicent Estruch and C{\'e}sar Ferri and Jos{\'e} Hern{\'a}ndez-Orallo and M. Jos{\'e} Ram\'{\i}rez-Quintana}, title = {Bagging Decision Multi-trees.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {41-51}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=41}, } @INPROCEEDINGS{GunterB04, author = {Simon G{\"u}nter and Horst Bunke}, title = {Ensembles of Classifiers Derived from Multiple Prototypes and Their Application to Handwriting Recognition.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {314-323}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=314}, } @INPROCEEDINGS{Hernandez-EspinosaFT04, author = {Carlos Hern{\'a}ndez-Espinosa and Mercedes Fern{\'a}ndez-Redondo and Joaqu\'{\i}n Torres-Sospedra}, title = {First Experiments on Ensembles of Radial Basis Functions.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {253-262}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=253}, } @INPROCEEDINGS{JuszczakD04, author = {Piotr Juszczak and Robert P. W. Duin}, title = {Combining One-Class Classifiers to Classify Missing Data.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {92-101}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=92}, } @INPROCEEDINGS{Kang04, author = {Hee-Joong Kang}, title = {Combining Classifiers Using Dependency-Based Product Approximation with Bayes Error Rate.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {112-121}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=112}, } @INPROCEEDINGS{KittlerS04, author = {Josef Kittler and Mohammad Sadeghi}, title = {Physics-Based Decorrelation of Image Data for Decision Level Fusion in Face Verification.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {354-363}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=354}, } @INPROCEEDINGS{Kuncheva04, author = {Ludmila I. Kuncheva}, title = {Classifier Ensembles for Changing Environments.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {1-15}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=1}, } @INPROCEEDINGS{MarcialisR04, author = {Gian Luca Marcialis and Fabio Roli}, title = {High Security Fingerprint Verification by Perceptron-Based Fusion of Multiple Matchers.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {364-373}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=364}, } @INPROCEEDINGS{MarroccoT04, author = {Claudio Marrocco and Francesco Tortorella}, title = {A Method for Designing Cost-Sensitive ECOC.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {204-213}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=204}, } @INPROCEEDINGS{MelvilleSMM04, author = {Prem Melville and Nishit Shah and Lilyana Mihalkova and Raymond J. Mooney}, title = {Experiments on Ensembles with Missing and Noisy Data.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {293-302}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=293}, } @INPROCEEDINGS{MuhlbaierTP04, author = {Michael Muhlbaier and Apostolos Topalis and Robi Polikar}, title = {Learn++.MT: A New Approach to Incremental Learning.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {52-61}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=52}, } @INPROCEEDINGS{Oza04, author = {Nikunj C. Oza}, title = {AveBoost2: Boosting for Noisy Data.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {31-40}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=31}, } @INPROCEEDINGS{PekalskaSD04, author = {Elzbieta Pekalska and Marina Skurichina and Robert P. W. Duin}, title = {Combining Dissimilarity-Based One-Class Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {122-133}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=122}, } @INPROCEEDINGS{RahmanTKA04, author = {Fuad Rahman and Yuliya Tarnikova and Aman Kumar and Hassan Alam}, title = {Second Guessing a Commercial 'Black Box' Classifier by an 'In House' Classifier: Serial Classifier Combination in a Speech Recognition Application.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {374-383}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=374}, } @INPROCEEDINGS{RajanG04, author = {Suju Rajan and Joydeep Ghosh}, title = {An Empirical Comparison of Hierarchical vs. Two-Level Approaches to Multiclass Problems.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {283-292}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=283}, } @INPROCEEDINGS{Rao04, author = {Nageswara S. V. Rao}, title = {A Generic Sensor Fusion Problem: Classification and Function Estimation.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {16-30}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=16}, } @INPROCEEDINGS{RaudysI04, author = {Sarunas Raudys and Masakazu Iwamura}, title = {Multiple Classifiers System for Reducing Influences of Atypical Observations.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {233-242}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=233}, } @PROCEEDINGS{mcs2004, editor = {Fabio Roli and Josef Kittler and Terry Windeatt}, title = {Multiple Classifier Systems, 5th International Workshop, MCS 2004, Cagliari, Italy, June 9-11, 2004, Proceedings}, booktitle = {Multiple Classifier Systems}, publisher = {Springer}, series = {Lecture Notes in Computer Science}, volume = 3077, year = 2004, isbn = {3-540-22144-1}, bibsource = {DBLP, http://dblp.uni-trier.de}, } @INPROCEEDINGS{RooneyPAT04, author = {Niall Rooney and David W. Patterson and Sarab S. Anand and Alexey Tsymbal}, title = {Dynamic Integration of Regression Models.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {164-173}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=164}, } @INPROCEEDINGS{SaerensF04, author = {Marco Saerens and Francois Fouss}, title = {Yet Another Method for Combining Classifiers Outputs: A Maximum Entropy Approach.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {82-91}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=82}, } @INPROCEEDINGS{SchenkerBLK04, author = {Adam Schenker and Horst Bunke and Mark Last and Abraham Kandel}, title = {Building Graph-Based Classifier Ensembles by Random Node Selection.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {214-222}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=214}, } @INPROCEEDINGS{SvetnikLTW04, author = {Vladimir Svetnik and Andy Liaw and Christopher Tong and Ting Wang}, title = {Application of Breiman's Random Forest to Modeling Structure-Activity Relationships of Pharmaceutical Molecules.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {334-343}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=334}, } @INPROCEEDINGS{TapiaGHG04, author = {Elizabeth Tapia and Jos{\'e} Carlos Gonz{\'a}lez and Alexander H{\"u}termann and L. Javier Garc\'{\i}a-Villalba}, title = {Beyond Boosting: Recursive ECOC Learning Machines.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {62-71}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=62}, } @INPROCEEDINGS{Valentini04, author = {Giorgio Valentini}, title = {Random Aggregated and Bagged Ensembles of SVMs: An Empirical Bias?Variance Analysis.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {263-272}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=263}, } @INPROCEEDINGS{WangT04, author = {Xiaogang Wang and Xiaoou Tang}, title = {Experimental Study on Multiple LDA Classifier Combination for High Dimensional Data Classification.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {344-353}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=344}, } @INPROCEEDINGS{Windeatt04, author = {Terry Windeatt}, title = {Spectral Measure for Multi-class Problems.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {184-193}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=184}, } @INPROCEEDINGS{WindridgeB04, author = {David Windridge and Richard Bowden}, title = {Induced Decision Fusion in Automated Sign Language Interpretation: Using ICA to Isolate the Underlying Components of Sign.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {303-313}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=303}, } @INPROCEEDINGS{WindridgePK04, author = {David Windridge and Robin Patenall and Josef Kittler}, title = {The Relationship between Classifier Factorisation and Performance in Stochastic Vector Quantisation.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {194-203}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=194}, } @INPROCEEDINGS{ZouariHLA04, author = {H{\'e}la Zouari and Laurent Heutte and Yves Lecourtier and Adel M. Alimi}, title = {Building Diverse Classifier Outputs to Evaluate the Behavior of Combination Methods: The Case of Two Classifiers.}, booktitle = {Multiple Classifier Systems}, year = 2004, pages = {273-282}, ee = {http://springerlink.metapress.com/openurl.asp?genre=article{\&}issn=0302-9743{\&}volume=3077{\&}spage=273}, } #MCS05 @INPROCEEDINGS{AyadK05, author = {Hanan Ayad and Mohamed S. Kamel}, title = {Cluster-Based Cumulative Ensembles.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {236-245}, ee = {http://dx.doi.org/10.1007/11494683_24}, } @INPROCEEDINGS{BanfieldHBK05, author = {Robert E. Banfield and Lawrence O. Hall and Kevin W. Bowyer and W. Philip Kegelmeyer}, title = {Ensembles of Classifiers from Spatially Disjoint Data.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {196-205}, ee = {http://dx.doi.org/10.1007/11494683_20}, } @INPROCEEDINGS{BonissoneEG05, author = {Piero P. Bonissone and Neil Eklund and Kai Goebel}, title = {Using an Ensemble of Classifiers to Audit a Production Classifier.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {376-386}, ee = {http://dx.doi.org/10.1007/11494683_38}, } @INPROCEEDINGS{BrownWS05, author = {Gavin Brown and Jeremy L. Wyatt and Ping Sun}, title = {Between Two Extremes: Examining Decompositions of the Ensemble Objective Function.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {296-305}, ee = {http://dx.doi.org/10.1007/11494683_30}, } @INPROCEEDINGS{ChawlaB05, author = {Nitesh V. Chawla and Kevin W. Bowyer}, title = {Designing Multiple Classifier Systems for Face Recognition.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {407-416}, ee = {http://dx.doi.org/10.1007/11494683_41}, } @INPROCEEDINGS{ChenK05, author = {Lei Chen and Mohamed S. Kamel}, title = {Design of Multiple Classifier Systems for Time Series Data.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {216-225}, ee = {http://dx.doi.org/10.1007/11494683_22}, } @INPROCEEDINGS{ChindaroSF05, author = {Samuel Chindaro and Konstantinos Sirlantzis and Michael C. Fairhurst}, title = {Analysis and Modelling of Diversity Contribution to Ensemble-Based Texture Recognition Performance.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {387-396}, ee = {http://dx.doi.org/10.1007/11494683_39}, } @INPROCEEDINGS{DaraMK05, author = {Rozita A. Dara and Masoud Makrehchi and Mohamed S. Kamel}, title = {Data Partitioning Evaluation Measures for Classifier Ensembles.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {306-315}, ee = {http://dx.doi.org/10.1007/11494683_31}, } @INPROCEEDINGS{DuanK05, author = {Kai-Bo Duan and S. Sathiya Keerthi}, title = {Which Is the Best Multiclass SVM Method? An Empirical Study.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {278-285}, ee = {http://dx.doi.org/10.1007/11494683_28}, } @INPROCEEDINGS{ErdemPGY05, author = {Zeki Erdem and Robi Polikar and Fikret S. G{\"u}rgen and Nejat Yumusak}, title = {Ensemble of SVMs for Incremental Learning.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {246-256}, ee = {http://dx.doi.org/10.1007/11494683_25}, } @INPROCEEDINGS{ErdoganEEBEKA05, author = {Hakan Erdogan and Aytal Er\c{c}il and Hazim Kemal Ekenel and S. Y. Bilgin and Ibrahim Eden and Meltem Kiris\c{c}i and Huseyin Abut}, title = {Multi-modal Person Recognition for Vehicular Applications.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {366-375}, ee = {http://dx.doi.org/10.1007/11494683_37}, } @INPROCEEDINGS{Fierrez-AguilarGOG05, author = {Julian Fi{\'e}rrez-Aguilar and Daniel Garcia-Romero and Javier Ortega-Garcia and Joaquin Gonzalez-Rodriguez}, title = {Speaker Verification Using Adapted User-Dependent Multilevel Fusion.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {356-365}, ee = {http://dx.doi.org/10.1007/11494683_36}, } @INPROCEEDINGS{FumeraRS05, author = {Giorgio Fumera and Fabio Roli and Alessandra Serrau}, title = {Dynamics of Variance Reduction in Bagging and Other Techniques Based on Randomisation.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {316-325}, ee = {http://dx.doi.org/10.1007/11494683_32}, } @INPROCEEDINGS{Gal-OrMS05, author = {Mordechai Gal-Or and Jerrold H. May and William E. Spangler}, title = {Using Decision Tree Models and Diversity Measures in the Selection of Ensemble Classification Models.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {186-195}, ee = {http://dx.doi.org/10.1007/11494683_19}, } @INPROCEEDINGS{HessKK05, author = {Andreas He{\ss} and Rinat Khoussainov and Nicholas Kushmerick}, title = {Ensemble Learning with Biased Classifiers: The Triskel Algorithm.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {226-235}, ee = {http://dx.doi.org/10.1007/11494683_23}, } @INPROCEEDINGS{KapoorAP05, author = {Ashish Kapoor and Hyungil Ahn and Rosalind W. Picard}, title = {Mixture of Gaussian Processes for Combining Multiple Modalities.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {86-96}, ee = {http://dx.doi.org/10.1007/11494683_9}, } @INPROCEEDINGS{KimK05, author = {Eunju Kim and Jaepil Ko}, title = {Dynamic Classifier Integration Method.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {97-107}, ee = {http://dx.doi.org/10.1007/11494683_10}, } @INPROCEEDINGS{LandgrebePTD05, author = {Thomas Landgrebe and Pavel Pacl\'{\i}k and David M. J. Tax and Robert P. W. Duin}, title = {Optimising Two-Stage Recognition Systems.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {206-215}, ee = {http://dx.doi.org/10.1007/11494683_21}, } @INPROCEEDINGS{LeiG05, author = {Hansheng Lei and Venu Govindaraju}, title = {Half-Against-Half Multi-class Support Vector Machines.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {156-164}, ee = {http://dx.doi.org/10.1007/11494683_16}, } @INPROCEEDINGS{LiCFK05, author = {Peng Li and Kap Luk Chan and Sheng Fu and Shankar Muthu Krishnan}, title = {An Abnormal ECG Beat Detection Approach for Long-Term Monitoring of Heart Patients Based on Hybrid Kernel Machine Ensemble.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {346-355}, ee = {http://dx.doi.org/10.1007/11494683_35}, } @inproceedings{MelnikVZ05, author = {Ofer Melnik and Yehuda Vardi and Cun-Hui Zhang}, title = {A Probability Model for Combining Ranks.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {64-73}, ee = {http://dx.doi.org/10.1007/11494683_7}, } @INPROCEEDINGS{MuhlbaierTP05, author = {Michael Muhlbaier and Apostolos Topalis and Robi Polikar}, title = {Ensemble Confidence Estimates Posterior Probability.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {326-335}, ee = {http://dx.doi.org/10.1007/11494683_33}, } @INPROCEEDINGS{Narasimhamurthy05, author = {Anand M. Narasimhamurthy}, title = {Evaluation of Diversity Measures for Binary Classifier Ensembles.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {267-277}, ee = {http://dx.doi.org/10.1007/11494683_27}, } @INPROCEEDINGS{NishidaK05, author = {Kenji Nishida and Takio Kurita}, title = {Boosting Soft-Margin SVM with Feature Selection for Pedestrian Detection.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {22-31}, ee = {http://dx.doi.org/10.1007/11494683_3}, } @INPROCEEDINGS{NishidaYO05, author = {Kyosuke Nishida and Koichiro Yamauchi and Takashi Omori}, title = {ACE: Adaptive Classifiers-Ensemble System for Concept-Drifting Environments.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {176-185}, ee = {http://dx.doi.org/10.1007/11494683_18}, } @INPROCEEDINGS{PaclikLTD05, author = {Pavel Pacl\'{\i}k and Thomas Landgrebe and David M. J. Tax and Robert P. W. Duin}, title = {On Deriving the Second-Stage Training Set for Trainable Combiners.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {136-146}, ee = {http://dx.doi.org/10.1007/11494683_14}, } @INPROCEEDINGS{PatenallWK05, author = {Robin Patenall and David Windridge and Josef Kittler}, title = {Multiple Classifier Fusion Performance in Networked Stochastic Vector Quantisers.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {128-135}, ee = {http://dx.doi.org/10.1007/11494683_13}, } @INPROCEEDINGS{PohB05, author = {Norman Poh and Samy Bengio}, title = {EER of Fixed and Trainable Fusion Classifiers: A Theoretical Study with Application to Biometric Authentication Tasks.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {74-85}, ee = {http://dx.doi.org/10.1007/11494683_8}, } @INPROCEEDINGS{PranckevicieneBS05, author = {Erinija Pranckeviciene and Richard Baumgartner and Ray L. Somorjai}, title = {Using Domain Knowledge for in the Random Subspace Method: Application: Application to the Classification of Biomedical Spectra.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {336-345}, ee = {http://dx.doi.org/10.1007/11494683_34}, } @INPROCEEDINGS{PriorW05, author = {Matthew Prior and Terry Windeatt}, title = {Over-Fitting in Ensembles of Neural Network Classifiers Within ECOC Frameworks.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {286-295}, ee = {http://dx.doi.org/10.1007/11494683_29}, } @INPROCEEDINGS{RajanG05, author = {Suju Rajan and Joydeep Ghosh}, title = {Exploiting Class Hierarchies for Knowledge Transfer in Hyperspectral Data.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {417-427}, ee = {http://dx.doi.org/10.1007/11494683_42}, } @INPROCEEDINGS{RedpathL05, author = {D. B. Redpath and K. Lebart}, title = {Observations on Boosting Feature Selection.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {32-41}, ee = {http://dx.doi.org/10.1007/11494683_4}, } @INPROCEEDINGS{Roli05, author = {Fabio Roli}, title = {Semi-supervised Multiple Classifier Systems: Background and Research Directions.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {1-11}, ee = {http://dx.doi.org/10.1007/11494683_1}, } @INPROCEEDINGS{SantoPSV05, author = {Massimo De Santo and Gennaro Percannella and Carlo Sansone and Mario Vento}, title = {Combining Audio-Based and Video-Based Shot Classification Systems for News Videos Segmentation.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {397-406}, ee = {http://dx.doi.org/10.1007/11494683_40}, } @INPROCEEDINGS{SkurichinaD05, author = {Marina Skurichina and Robert P. W. Duin}, title = {Combining Feature Subsets in Feature Selection.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {165-175}, ee = {http://dx.doi.org/10.1007/11494683_17}, } @INPROCEEDINGS{SmithW05, author = {R. S. Smith and Terry Windeatt}, title = {Decoding Rules for Error Correcting Output Code Ensembles.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {53-63}, ee = {http://dx.doi.org/10.1007/11494683_6}, } @INPROCEEDINGS{TapiaSG05, author = {Elizabeth Tapia and Esteban Serra and Jos{\'e} Carlos Gonz{\'a}lez}, title = {Recursive ECOC for Microarray Data Classification.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {108-117}, ee = {http://dx.doi.org/10.1007/11494683_11}, } @INPROCEEDINGS{ThielSP05, author = {Christian Thiel and Friedhelm Schwenker and G{\"u}nther Palm}, title = {Using Dempster-Shafer Theory in MCF Systems to Reject Samples.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {118-127}, ee = {http://dx.doi.org/10.1007/11494683_12}, } @INPROCEEDINGS{TianYH05, author = {Qi Tian and Jie Yu and Thomas S. Huang}, title = {Boosting Multiple Classifiers Constructed by Hybrid Discriminant Analysis.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {42-52}, ee = {http://dx.doi.org/10.1007/11494683_5}, } @INPROCEEDINGS{TulyakovG05, author = {Sergey Tulyakov and Venu Govindaraju}, title = {Using Independence Assumption to Improve Multimodal Biometric Fusion.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {147-155}, ee = {http://dx.doi.org/10.1007/11494683_15}, } @INPROCEEDINGS{WangZL05, author = {Fei Wang and Changshui Zhang and Naijiang Lu}, title = {Boosting GMM and Its Two Applications.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {12-21}, ee = {http://dx.doi.org/10.1007/11494683_2}, } @INPROCEEDINGS{YangQ05, author = {Li-ying Yang and Zheng Qin}, title = {Design of a New Classifier Simulator.}, booktitle = {Multiple Classifier Systems}, year = 2005, pages = {257-266}, ee = {http://dx.doi.org/10.1007/11494683_26}, } #MCS07 @inproceedings{mottl04combining, author = "Vadim Mottl and Alexander Tatarchuk and Valentina Sulimova and Olga Krasotkina and Oleg Seredin", title = "Combining Pattern Recognition Modalities at the Sensor Level Via Kernel Fusion", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "1-12", year = "2007", address = "Prague, Czech Republic" } @inproceedings{windridge04neutral, author = "David Windridge and Vadim Mottl and Alexander Tatarchuk and Andrey Eliseyev", title = "The Neutral Point Method for Kernel-Based Combination of Disjoint Training Data in Multi-modal Pattern Recognition", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "13-21", year = "2007", address = "Prague, Czech Republic" } @inproceedings{lee04kernel, author = "Wan-Jui Lee and Sergey Verzakov and Robert P. W. Duin", title = "Kernel Combination Versus Classifier Combination", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "22-31", year = "2007", address = "Prague, Czech Republic" } @inproceedings{merler04deriving, author = "Stefano Merler and Giuseppe Jurman and Cesare Furlanello", title = "Deriving the Kernel from Training Data", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "32-41", year = "2007", address = "Prague, Czech Republic" } @inproceedings{lienemann04application, author = "Kai Lienemann and Thomas Plötz and Gernot A. Fink", title = "On the Application of SVM-Ensembles Based on Adapted Random Subspace Sampling for Automatic Classification of NMR Data", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "42-51", year = "2007", address = "Prague, Czech Republic" } @inproceedings{ko04new, author = "Albert Hung-Ren Ko and Robert Sabourin and Alceu de Souza Britto Jr.", title = "A New HMM-Based Ensemble Generation Method for Numeral Recognition", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "52-61", year = "2007", address = "Prague, Czech Republic" } @inproceedings{raudys04classifiers, author = "Sarunas Raudys and Ömer Kaan Baykan and Ahmet Babalik and Vitalij Denisov and Antanas Andrius Bielskis", title = "Classifiers Fusion in Recognition of Wheat Varieties", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "62-71", year = "2007", address = "Prague, Czech Republic" } @inproceedings{bertolami04multiple, author = "Roman Bertolami and Horst Bunke", title = "Multiple Classifier Methods for Offline Handwritten Text Line Recognition", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "72-81", year = "2007", address = "Prague, Czech Republic" } @inproceedings{christensen04applying, author = "Hans Ulrich Christensen and Daniel Ortiz Arroyo", title = "Applying Data Fusion Methods to Passage Retrieval in QAS", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "82-92", year = "2007", address = "Prague, Czech Republic" } @inproceedings{mohamed04co-training, author = "Tawfik A. Mohamed and Neamat El Gayar and Amir F. Atiya", title = "A Co-training Approach for Time Series Prediction with Missing Data", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "93-102", year = "2007", address = "Prague, Czech Republic" } @inproceedings{sun04improved, author = "Shiliang Sun", title = "An Improved Random Subspace Method and Its Application to EEG Signal Classification", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "103-112", year = "2007", address = "Prague, Czech Republic" } @inproceedings{sun04ensemble, author = "Shiliang Sun", title = "Ensemble Learning Methods for Classifying EEG Signals", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "113-120", year = "2007", address = "Prague, Czech Republic" } @inproceedings{sadeghi04confidence, author = "Mohammad Sadeghi and Samaneh Khoshrou and Josef Kittler", title = "Confidence Based Gating of Colour Features for Face Authentication", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "121-130", year = "2007", address = "Prague, Czech Republic" } @inproceedings{ebrahimpour04view-based, author = "Reza Ebrahimpour and Ehsanollah Kabir and Mohammad Reza Yousefi", title = "View-Based Eigenspaces with Mixture of Experts for View-Independent Face Recognition", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "131-140", year = "2007", address = "Prague, Czech Republic" } @inproceedings{serrano04fusion, author = "Ángel Serrano and Isaac Martín de Diego and Cristina Conde and Enrique Cabello and Li Bai and LinLin Shen", title = "Fusion of Support Vector Classifiers for Parallel Gabor Methods Applied to Face Verification", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "141-150", year = "2007", address = "Prague, Czech Republic" } @inproceedings{marcialis04serial, author = "Gian Luca Marcialis and Fabio Roli", title = "Serial Fusion of Fingerprint and Face Matchers", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "151-160", year = "2007", address = "Prague, Czech Republic" } @inproceedings{hall04boosting, author = "Lawrence O. Hall and Robert E. Banfield and Kevin W. Bowyer and W. Philip Kegelmeyer", title = "Boosting Lite - Handling Larger Datasets and Slower Base Classifiers", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "161-170", year = "2007", address = "Prague, Czech Republic" } @inproceedings{meynet04information, author = "Julien Meynet and Jean-Philippe Thiran", title = "Information Theoretic Combination of Classifiers with Application to AdaBoost", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "171-179", year = "2007", address = "Prague, Czech Republic" } @inproceedings{lu04interactive, author = "Yijuan Lu and Qi Tian and Thomas S. Huang", title = "Interactive Boosting for Image Classification", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "180-189", year = "2007", address = "Prague, Czech Republic" } @inproceedings{bicego04group-induced, author = "Manuele Bicego and Elzbieta Pekalska and Robert P. W. Duin", title = "Group-Induced Vector Spaces", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "190-199", year = "2007", address = "Prague, Czech Republic" } @inproceedings{hadjitodorov04selecting, author = "Stefan Todorov Hadjitodorov and Ludmila I. Kuncheva", title = "Selecting Diversifying Heuristics for Cluster Ensembles", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "200-209", year = "2007", address = "Prague, Czech Republic" } @inproceedings{haindl04unsupervised, author = "Michal Haindl and Stanislav Mikes", title = "Unsupervised Texture Segmentation Using Multiple Segmenters Strategy", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "210-219", year = "2007", address = "Prague, Czech Republic" } @inproceedings{riesen04classifier, author = "Kaspar Riesen and Horst Bunke", title = "Classifier Ensembles for Vector Space Embedding of Graphs", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "220-230", year = "2007", address = "Prague, Czech Republic" } @inproceedings{maudes04cascading, author = "Jesús Maudes and Juan J. Rodríguez and Cesar García-Osorio", title = "Cascading for Nominal Data", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "231-240", year = "2007", address = "Prague, Czech Republic" } @inproceedings{kudo04combination, author = "Mineichi Kudo and Satoshi Shirai and Hiroshi Tenmoto", title = "A Combination of Sample Subsets and Feature Subsets in One-Against-Other Classifiers", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "241-250", year = "2007", address = "Prague, Czech Republic" } @inproceedings{depasquale04random, author = "Joseph DePasquale and Robi Polikar", title = "Random Feature Subset Selection for Ensemble Based Classification of Data with Missing Features", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "251-260", year = "2007", address = "Prague, Czech Republic" } @inproceedings{silva04feature, author = "Hugo Silva and Ana L. N. Fred", title = "Feature Subspace Ensembles: A Parallel Classifier Combination Scheme Using Feature Selection", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "261-270", year = "2007", address = "Prague, Czech Republic" } @inproceedings{windeatt04stopping, author = "Terry Windeatt and Matthew Prior", title = "Stopping Criteria for Ensemble-Based Feature Selection", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "271-281", year = "2007", address = "Prague, Czech Republic" } @inproceedings{foggia04rejecting, author = "Pasquale Foggia and Gennaro Percannella and Carlo Sansone and Mario Vento", title = "On Rejecting Unreliably Classified Patterns", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "282-291", year = "2007", address = "Prague, Czech Republic" } @inproceedings{biggio04bayesian, author = "Battista Biggio and Giorgio Fumera and Fabio Roli", title = "Bayesian Analysis of Linear Combiners", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "292-301", year = "2007", address = "Prague, Czech Republic" } @inproceedings{ko04applying, author = "Albert Hung-Ren Ko and Robert Sabourin and Alceu de Souza Britto Jr.", title = "Applying Pairwise Fusion Matrix on Fusion Functions for Classifier Combination", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "302-311", year = "2007", address = "Prague, Czech Republic" } @inproceedings{chindaro04modelling, author = "Samuel Chindaro and Konstantinos Sirlantzis and Michael C. Fairhurst", title = "Modelling Multiple-Classifier Relationships Using Bayesian Belief Networks", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "312-321", year = "2007", address = "Prague, Czech Republic" } @inproceedings{li04classifier, author = "Shoushan Li and Chengqing Zong", title = "Classifier Combining Rules Under Independence Assumptions", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "322-332", year = "2007", address = "Prague, Czech Republic" } @inproceedings{marrocco04embedding, author = "Claudio Marrocco and Paolo Simeone and Francesco Tortorella", title = "Embedding Reject Option in ECOC Through LDPC Codes", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "333-343", year = "2007", address = "Prague, Czech Republic" } @inproceedings{poh04combination, author = "Norman Poh and Guillaume Heusch and Josef Kittler", title = "On Combination of Face Authentication Experts by a Mixture of Quality Dependent Fusion Classifiers", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "344-356", year = "2007", address = "Prague, Czech Republic" } @inproceedings{tronci04index, author = "Roberto Tronci and Giorgio Giacinto and Fabio Roli", title = "Index Driven Combination of Multiple Biometric Experts for AUC Maximisation", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "357-366", year = "2007", address = "Prague, Czech Republic" } @inproceedings{kryszczuk04q-stack:, author = "Krzysztof Kryszczuk and Andrzej Drygajlo", title = "Q-stack: Uni- and Multimodal Classifier Stacking with Quality Measures", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "367-376", year = "2007", address = "Prague, Czech Republic" } @inproceedings{richiardi04reliability-based, author = "Jonas Richiardi and Andrzej Drygajlo", title = "Reliability-Based Voting Schemes Using Modality-Independent Features in Multi-classifier Biometric Authentication", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "377-386", year = "2007", address = "Prague, Czech Republic" } @inproceedings{tulyakov04optimal, author = "Sergey Tulyakov and Venu Govindaraju and Chaohong Wu", title = "Optimal Classifier Combination Rules for Verification and Identification Systems", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "387-396", year = "2007", address = "Prague, Czech Republic" } @inproceedings{chawla04exploiting, author = "Nitesh V. Chawla and Jared Sylvester", title = "Exploiting Diversity in Ensembles: Improving the Performance on Unbalanced Datasets", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "397-406", year = "2007", address = "Prague, Czech Republic" } @inproceedings{chung04diversity-performance, author = "Yun Sheng Chung and D. Frank Hsu and Chuan Yi Tang", title = "On the Diversity-Performance Relationship for Majority Voting in Classifier Ensembles", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "407-420", year = "2007", address = "Prague, Czech Republic" } @inproceedings{cecotti04hierarchical, author = "Hubert Cecotti and Abdel Belaïd", title = "Hierarchical Behavior Knowledge Space", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "421-430", year = "2007", address = "Prague, Czech Republic" } @inproceedings{ko04new, author = "Albert Hung-Ren Ko and Robert Sabourin and Alceu de Souza Britto Jr.", title = "A New Dynamic Ensemble Selection Method for Numeral Recognition", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "431-439", year = "2007", address = "Prague, Czech Republic" } @inproceedings{zanda04ensemble, author = "Manuela Zanda and Gavin Brown and Giorgio Fumera and Fabio Roli", title = "Ensemble Learning in Linearly Combined Classifiers Via Negative Correlation", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "440-449", year = "2007", address = "Prague, Czech Republic" } @inproceedings{rodríguez04naïve, author = "Juan J. Rodríguez and Ludmila I. Kuncheva", title = "Naïve Bayes Ensembles with a Random Oracle", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "450-458", year = "2007", address = "Prague, Czech Republic" } @inproceedings{kuncheva04experimental, author = "Ludmila I. Kuncheva and Juan J. Rodríguez", title = "An Experimental Study on Rotation Forest Ensembles", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "459-468", year = "2007", address = "Prague, Czech Republic" } @inproceedings{kanevskiy04cooperative, author = "Daniel Kanevskiy and Konstantin Vorontsov", title = "Cooperative Coevolutionary Ensemble Learning", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "469-478", year = "2007", address = "Prague, Czech Republic" } @inproceedings{berkman04robust, author = "Omer Berkman and Nathan Intrator", title = "Robust Inference in Bayesian Networks with Application to Gene Expression Temporal Data", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "479-489", year = "2007", address = "Prague, Czech Republic" } @inproceedings{muhlbaier04ensemble, author = "Michael Muhlbaier and Robi Polikar", title = "An Ensemble Approach for Incremental Learning in Nonstationary Environments", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "490-500", year = "2007", address = "Prague, Czech Republic" } @inproceedings{benediktsson04multiple, author = "Jon Atli Benediktsson and Jocelyn Chanussot and Mathieu Fauvel", title = "Multiple Classifier Systems in Remote Sensing: From Basics to Recent Developments", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "501-512", year = "2007", address = "Prague, Czech Republic" } @inproceedings{bengio04biometric, author = "Samy Bengio and Johnny Marithoz", title = "Biometric Person Authentication Is a Multiple Classifier Problem", editor = "M. Haindl, J. Kittler, F. Roli", booktitle = "Proc. Int. Workshop on Multiple Classifier Systems (LNCS 4472)", publisher = "Springer", month = "May 23-25", pages = "513-522", year = "2007", address = "Prague, Czech Republic" } #POST2000NONMCSPAPERS @article{martinez06using, author = {Gonzalo Mart{\'{\i}}nez-Mu{\~n}oz and Alberto Su{\'a}rez}, title = {Using Boosting to Prune Bagging Ensembles}, journal = {Pattern Recognition Letters}, pages = {156--165}, volume = {28}, year = {2007}, number = {1}, abstract = {Boosting is used to determine the order in which classifiers are aggregated in a bagging ensemble. Early stopping in the aggregation of the classifiers in the ordered bagging ensemble allows the identification of subensembles that require less memory for storage, classify faster and can improve the generalization accuracy of the original bagging ensemble. In all the classification problems investigated pruned ensembles with 20\% of the original classifiers show statistically significant improvements over bagging. In problems where boosting is superior to bagging, these improvements are not sufficient to reach the accuracy of the corresponding boosting ensembles. However, ensemble pruning preserves the performance of bagging in noisy classification tasks, where boosting often has larger generalization errors. Therefore, pruned bagging should generally be preferred to complete bagging and, if no information about the level of noise is available, it is a robust alternative to AdaBoost.} } @article{martinez05switching, author = {Martinez-Munoz, G. and Suarez, A.}, title = {Switching class labels to generate classification ensembles}, journal = {Pattern Recognition}, year = {2005}, volume = {38}, pages = {1483--1494}, number = {10}, abstract = {Ensembles that combine the decisions of classifiers generated by using perturbed versions of the training set where the classes of the training examples are randomly switched can produce a significant error reduction, provided that large numbers of units and high class switching rates are used. The classifiers generated by this procedure have statistically uncorrelated errors in the training set. Hence, the ensembles they form exhibit a similar dependence of the training error on ensemble size, independently of the classification problem. In particular, for binary classification problems, the classification performance of the ensemble on the training data can be analysed in terms of a Bernoulli process. Experiments on several UCI datasets demonstrate the improvements in classification accuracy that can be obtained using these class-switching ensembles. (c) 2005 Pattern Recognition Society. Published by Elsevier Ltd. All rights reserved.} } @article{martinez04using, author = {Gonzalo Mart{\'{\i}}nez-Mu{\~n}oz and Alberto Su{\'a}rez}, title = {Using All Data to Generate Decision Tree Ensembles}, journal = {IEEE Transactions on Systems, Man and Cybernetics part {C}}, year = {2004}, volume = {34}, pages = {393--397}, number = {4} } @inproceedings{hernandez06pruning, author = {Daniel Hern{\'a}ndez-Lobato and Jos{\'e} Miguel Hern{\'a}ndez-Lobato and Rub{\'e}n Ruiz-Torrubiano and {\'A}ngel Valle}, title = {Pruning Adaptive Boosting Ensembles by Means of a Genetic Algorithm.}, booktitle = {IDEAL}, year = {2006}, pages = {322--329} } @inproceedings{martinez06building, author = {Gonzalo Mart\'{\i}nez-Mu\~noz and Aitor S\'anchez-Mart\'{\i}nez and Daniel Hern\'andez-Lobato and Alberto Su\'arez} title = {Building Ensembles of Neural Networks with Class-Switching}, booktitle = {Artificial Neural Networks - ICANN 2006}, pages = {178--187}, year = {2006} } @inproceedings{martinez06pruning, author = {Gonzalo Mart\'{\i}nez-Mu\~{n}oz and Alberto Su\'{a}rez}, title = {Pruning in ordered bagging ensembles}, booktitle = {ICML '06: Proceedings of the 23rd international conference on Machine learning}, year = {2006}, isbn = {1-59593-383-2}, pages = {609--616}, location = {Pittsburgh, Pennsylvania}, doi = {http://doi.acm.org/10.1145/1143844.1143921}, publisher = {ACM Press}, address = {New York, NY, USA}, abstract = {We present a novel ensemble pruning method based on reordering the classifiers obtained from bagging and then selecting a subset for aggregation. Ordering the classifiers generated in bagging makes it possible to build subensembles of increasing size by including first those classifiers that are expected to perform best when aggregated. Ensemble pruning is achieved by halting the aggregation process before all the classifiers generated are included into the ensemble. Pruned subensembles containing between 15\% and 30\% of the initial pool of classifiers, besides being smaller, improve the generalization performance of the full bagging ensemble in the classification problems investigated.} } @inproceedings{hernandez06pruning, title = {Pruning in Ordered Regression Bagging Ensembles}, author = {Hern\'andez-Lobato, D.; Mart\'{\i}nez-Mu\~noz, G.; Su\'arez, A.}, booktitle = {Neural Networks, 2006. IJCNN '06. International Joint Conference on}, year = {2006}, pages = {1266--1273}, abstract = {An efficient procedure for pruning regression ensembles is introduced. Starting from a bagging ensemble, pruning proceeds by ordering the regressors in the original ensemble and then selecting a subset for aggregation. Ensembles of increasing size are built by including first the regressors that perform best when aggregated. This strategy gives an approximate solution to the problem of extracting from the original ensemble the minimum error subensemble, which we prove to be NP-hard. Experiments show that pruned ensembles with only 20\% of the initial regressors achieve better generalization accuracies than the complete bagging ensembles. The performance of pruned ensembles is analyzed by means of the bias-variance decomposition of the error.} } @inproceedings{martinez04aggregation, author = {Gonzalo Mart{\'{\i}}nez-Mu{\~n}oz and Alberto Su{\'a}rez}, title = {Aggregation Ordering in Bagging}, booktitle = {Proc. of the {IASTED} International Conference on Artificial Intelligence and Applications}, year = {2004}, pages = {258--263}, publisher = {Acta Press} } @inproceedings{Dzeroski02Combining, author = {Sa{\v{s}}o D{\v{z}}eroski and Bernard {\v{Z}}enko}, title = {Is Combining Classifiers Better than Selecting the Best One}, booktitle = {ICML '02: Proceedings of the Nineteenth International Conference on Machine Learning}, year = {2002}, isbn = {1-55860-873-7}, pages = {123--130}, publisher = {Morgan Kaufmann}, address = {San Francisco, CA, USA}, url = {http://www-ai.ijs.si/SasoDzeroski/files/2002_DZ_CombiningClassifiersBetter.pdf}, } @inproceedings{Zenko02Stacking, author = {Bernard {\v{Z}}enko and Sa{\v{s}}o D{\v{z}}eroski}, title = {{Stacking with an Extended Set of Meta-level Attributes and MLR}}, booktitle = {ECML '02: Proceedings of the 13th European Conference on Machine Learning}, year = {2002}, isbn = {3-540-44036-4}, pages = {493--504}, publisher = {Springer}, address = {Berlin, Germany}, url = {http://www.springerlink.com/content/qnr9tdtae9n27g3v/}, } @article{Dzeroski04Combining, author = {Sa{\v{s}}o D{\v{z}}eroski and Bernard {\v{Z}}enko}, title = {{Is Combining Classifiers with Stacking Better than Selecting the Best One?}}, journal = {Machine Learning}, volume = {54}, number = {3}, year = {2004}, issn = {0885-6125}, pages = {255--273}, doi = {http://dx.doi.org/10.1023/B:MACH.0000015881.36452.6e}, publisher = {Kluwer}, address = {Hingham, MA, USA}, } @article{rodriguez2006rfn, title={{Rotation forest: A new classifier ensemble method}}, author={RODRIGUEZ, J.J. and KUNCHEVA, L.I. and ALONSO, C.J.}, journal={IEEE transactions on pattern analysis and machine intelligence}, volume={28}, number={10}, pages={1619--1630}, year={2006}, publisher={Institute of Electrical and Electronics Engineers} } @article{windeatt2006ada, title={{Accuracy/Diversity and Ensemble MLP Classifier Design}}, author={Windeatt, T.}, journal={Neural Networks, IEEE Transactions on}, volume={17}, number={5}, pages={1194--1211}, year={2006} } @article{ko2006nof, title={{A New Objective Function for Ensemble Selection in Random Subspaces}}, author={Ko, A.H.R. and Sabourin, R. and de Souza Britto Jr, A.}, journal={Proceedings of the 18th International Conference on Pattern Recognition (ICPR'06)-Volume 04}, pages={185--188}, year={2006}, publisher={IEEE Computer Society Washington, DC, USA} } @article{tang2006adm, title={{An analysis of diversity measures}}, author={Tang, E.K. and Suganthan, P.N. and Yao, X.}, journal={Machine Learning}, volume={65}, number={1}, pages={247--271}, year={2006}, publisher={Springer} } @article{canuto2006uwd, title={{Using weighted dynamic classifier selection methods in ensembles with different levels of diversity}}, author={Canuto, A.M.P. and Fagundes, D. and Abreu, M.C.C. and Junior, J.C.X.}, journal={International Journal of Hybrid Intelligent Systems}, volume={3}, number={3}, pages={147--158}, year={2006}, publisher={IOS Press} } @article{kotsiantis2006lah, title={{Local averaging of heterogeneous regression models}}, author={Kotsiantis, SB}, journal={International Journal of Hybrid Intelligent Systems}, volume={3}, number={2}, pages={99--107}, year={2006}, publisher={IOS Press} } @article{zhang2006epv, title={{Ensemble Pruning Via Semi-definite Programming}}, author={Zhang, Y. and Burer, S. and Street, W.N.}, journal={Journal of Machine Learning Research}, volume={7}, pages={1315--1338}, year={2006} } @article{prior2006ptu, title={{Parameter Tuning using the Out-of-Bootstrap Generalisation Error Estimate for Stochastic Discrimination and Random Forests}}, author={Prior, M. and Windeatt, T.}, journal={Proceedings of the 18th International Conference on Pattern Recognition (ICPR'06)-Volume 02}, pages={498--501}, year={2006}, publisher={IEEE Computer Society Washington, DC, USA} } @article{lefaucheur2006rme, title={{Robust Multiclass Ensemble Classifiers via Symmetric Functions}}, author={Lefaucheur, P. and Nock, R.}, journal={Proceedings of the 18th International Conference on Pattern Recognition (ICPR'06)-Volume 04}, pages={136--139}, year={2006}, publisher={IEEE Computer Society Washington, DC, USA} } @article{ko2006eec, title={{Evolving ensemble of classifiers in random subspace}}, author={Ko, A.H.R. and Sabourin, R. and de Souza Britto Jr, A.}, journal={Proceedings of the 8th annual conference on Genetic and evolutionary computation}, pages={1473--1480}, year={2006}, publisher={ACM Press New York, NY, USA} } @article{kim2006oec, title={{Optimal ensemble construction via meta-evolutionary ensembles}}, author={Kim, Y.S. and Street, W.N. and Menczer, F.}, journal={Expert Systems With Applications}, volume={30}, number={4}, pages={705--714}, year={2006}, publisher={Elsevier} } @article{karmaker2006bar, title={{A boosting approach to remove class label noise}}, author={Karmaker, A. and Kwek, S.}, journal={International Journal of Hybrid Intelligent Systems}, volume={3}, number={3}, pages={169--177}, year={2006}, publisher={IOS Press} } @article{chandra2006elu, title={{Ensemble Learning Using Multi-Objective Evolutionary Algorithms}}, author={Chandra, A. and Yao, X.}, journal={Journal of Mathematical Modelling and Algorithms}, volume={5}, number={4}, pages={417--445}, year={2006}, publisher={Springer} } @article{hadjitodorov2006mdb, title={{Moderate diversity for better cluster ensembles}}, author={Hadjitodorov, S.T. and Kuncheva, L.I. and Todorova, L.P.}, journal={Information Fusion}, volume={7}, number={3}, pages={264--275}, year={2006}, publisher={Elsevier} } @ARTICLE{atikorale03hong, author = {A. S. Atukorale, T. Downs and P. N. Suganthan}, title = {Boosting HONG Networks}, journal = {Neurocomputing}, year = 2003, volume = 51, pages = {75-86}, month = {April}, } @INPROCEEDINGS{ahn01speciated, author = {Joon-Hyun Ahn and Sung-Bae Cho}, title = {Speciated Neural Networks Evolved with Fitness Sharing Technique}, booktitle = {Proceedings of the Congress on Evolutionary Computation}, address = {Seoul, Korea}, month = {May 27-30}, pages = {390--396}, year = 2001, } @ARTICLE{anthony04lists, author = {Martin Anthony}, title = {Generalisation Error Bounds for Threshold Decision Lists}, journal = {Journal of {M}achine {L}earning {R}esearch}, year = 2004, volume = 5, pages = {189--217}, abstract = {In this paper we consider the generalization accuracy of classification methods based on the iterative use of linear classifiers. The resulting classifiers, which we call threshold decision lists act as follows. Some points of the data set to be classified are given a particular classification according to a linear threshold function (or hyperplane). These are then removed from consideration, and the procedure is iterated until all points are classified. Geometrically, we can imagine that at each stage, points of the same classification are successively chopped off from the data set by a hyperplane. We analyse theoretically the generalization properties of data classification techniques that are based on the use of threshold decision lists and on the special subclass of multilevel threshold functions. We present bounds on the generalization error in a standard probabilistic learning framework. The primary focus in this paper is on obtaining generalization error bounds that depend on the levels of separation---or margins---achieved by the successive linear classifiers. We also improve and extend previously published theoretical bounds on the generalization ability of perceptron decision trees.}, } @INPROCEEDINGS{bahler00, author = {Dennis Bahler and Laura Navarro}, title = {Methods for Combining Heterogeneous Sets of Classifiers}, booktitle = {17th Natl. Conf. on Artificial Intelligence (AAAI), Workshop on New Research Problems for Machine Learning}, year = 2000, } @article{banfield05thinning, author = {Robert E. Banfield and Lawrence O. Hall and Kevin W. Bowyer and W. Philip Kegelmeyer}, title = {Ensemble diversity measures and their application to thinning.}, journal = {Information Fusion}, volume = {6}, number = {1}, year = {2005}, pages = {49-62} } @ARTICLE{baldridge06active, title = {{Active Learning and Logarithmic Opinion Pools for HPSG Parse Selection}}, author = {Baldridge, J. and Osborne, M.}, journal = {Natural Language Engineering (in press)}, } @ARTICLE{baram04choice, author = {Y. Baram and R. El-Yaniv and K. Luz}, title = {Online Choice of Active Learning Algorithms}, journal = {Journal of Machine Learning Research}, year = 2004, volume = 5, pages = {255--291}, month = {March}, } @ARTICLE{brameier01evolving, author = {Markus Brameier and Wolfgang Banzhaf}, title = {Evolving Teams of Predictors with Linear Genetic Programming}, journal = {Genetic Programming and Evolvable Machines}, volume = 2, number = 4, pages = {381--407}, year = 2001, } @TECHREPORT{breiman00infinite, author = {L.~Breiman}, title = {Some Infinite Theory for Predictor Ensembles}, institution = {Statistics Department, UC Berkeley}, year = 2000, number = 577, month = {August}, pdf = {http://www.boosting.org/papers/some_theory2001.pdf}, } @PHDTHESIS{brown04thesis, author = {G. Brown}, title = {Diversity in Neural Network Ensembles}, school = {School of Computer Science, University of Birmingham}, year = 2004, } @INPROCEEDINGS{brownwyatt03ambiguity, author = {G. Brown and J.L. Wyatt}, title = {The Use of the Ambiguity Decomposition in Neural Network Ensemble Learning Methods}, booktitle = {20th {I}nternational {C}onference on {M}achine {L}earning (ICML'03)}, year = 2003, editor = {Tom Fawcett and Nina Mishra}, month = {August}, address = {Washington DC, USA}, } @ARTICLE{brown04survey, author = {G. Brown and J.L. Wyatt and R. Harris and X. Yao}, title = {Diversity Creation Methods: A Survey and Categorisation}, journal = {Journal of Information Fusion}, volume = 6, number = 1, month = {March}, pages = {5--20}, year = 2005, } @ARTICLE{brownwyatt05jmlr, title = {Managing Diversity in Regression Ensembles}, author = {G. Brown and J.L. Wyatt and P. Tino}, journal = {Journal of Machine Learning Research}, volume = 6, pages = {1621--1650}, year = 2005, } @INPROCEEDINGS{brownyao01:ukci, author = {Gavin Brown and Xin Yao}, title = {On The {E}ffectiveness of {N}egative {C}orrelation {L}earning}, booktitle = {Proceedings of First UK Workshop on Computational Intelligence}, note = {Edinburgh, Scotland}, pages = {57-62}, year = 2001, } @INPROCEEDINGS{brownyao02exploiting, author = {Gavin Brown and Xin Yao and Jeremy Wyatt and Heiko Wersing and Bernhard Sendhoff}, title = {Exploiting Ensemble Diversity for Automatic Feature Extraction}, booktitle = {Proc. of the 9th International Conference on Neural Information Processing (ICONIP'02)}, pages = {1786-1790}, year = 2002, month = {November}, } @TECHREPORT{buhlmann00explaining, author = {Peter Buhlmann and Bin Yu}, title = {Explaining Bagging}, institution = {ETH Zurich, Seminar Fur Statistik}, month = {May}, year = 2000, number = 92, url = {ftp://ftp.stat.math.ethz.ch/Research-Reports/92.html}, } @INPROCEEDINGS{caruana06icml, author = {Rich Caruana and Alexandru Niculescu-Mizil}, title = {An Empirical Comparison of Supervised Learning Algorithms}, booktitle = {International Conference on Machine Learning}, year = 2006, organization = {Department of Computer Science, Cornell University}, annote = {Large empirical comparison of ML methods. Boosting and Bagging come top.}, } @INPROCEEDINGS{chawla01small, author = {Nitesh Chawla and Thomas Moore and Kevin Bowyer and Lawrence Hall and Clayton Springer and Philip Kegelmeyer}, title = {Bagging is a Small-Data-Set Phenomenon}, booktitle = {International Conference on Computer Vision and Pattern Recognition (CVPR)}, year = 2001, } @INPROCEEDINGS{cheung03radial, author = {Y.M. Cheung and R.B. Huang}, title = {An Advance on Divide-and-Conquer Based Radial Basis Function Networks}, booktitle = {Proceedings of Fourth International Conference on Intelligent Data Engineering and Automated Learning}, address = {Hong Kong}, month = {March}, year = 2003, } @ARTICLE{cohenIntrator01, author = {S. Cohen and N. Intrator}, year = 2002, title = {A hybrid projection based and radial basis function architecture: Initial values and global optimization}, journal = {Pattern Anal. Appl. (Special issue on Fusion of Multiple Classifiers)}, volume = 5, number = 2, pages = {113--120}, } @ARTICLE{cohenIntratorFusion01, author = {S. Cohen and N. Intrator}, title = {Automatic model selection in a hybrid perceptron/radial network}, journal = {Information Fusion}, volume = 3, number = 4, pages = {259--266}, year = 2002, } @INPROCEEDINGS{cunningham00diversity, author = {Padraig Cunningham and John Carney}, title = {Diversity versus Quality in Classification Ensembles Based on Feature Selection}, booktitle = {LNCS - European Conference on Machine Learning}, volume = 1810, publisher = {Springer, Berlin}, pages = {109--116}, year = 2000, } @INPROCEEDINGS{davidson04stable, author = {Ian Davidson}, title = {An Ensemble Technique for Stable Learners with Performance Bounds}, booktitle = {Proceedings of the Nineteenth National Conference on Artificial Intelligence}, year = 2004, publisher = {AAAI Press}, pages = {300-335}, } @ARTICLE{dietterich00experimental, author = {Thomas G. Dietterich}, title = {An Experimental Comparison of Three Methods for Constructing Ensembles of Decision Trees: Bagging, Boosting, and Randomization}, journal = {Machine Learning}, volume = 40, number = 2, pages = {139-157}, year = 2000, } @INPROCEEDINGS{domingos00unified, author = {Pedro Domingos}, title = {A Unified Bias-Variance Decomposition and its Applications}, booktitle = {Proc. 17th International Conf. on Machine Learning}, publisher = {Morgan Kaufmann, San Francisco, CA}, pages = {231--238}, year = 2000, url = {http://citeseer.nj.nec.com/domingos00unified.html}, } @INPROCEEDINGS{domingos00unified-AAAI, author = {Pedro Domingos}, title = {A Unified Bias-Variance Decomposition for Zero-One and Squared Loss}, booktitle = {{AAAI}/{IAAI}}, pages = {564-569}, year = 2000, } @TECHREPORT{drug06formal, author = {Jan Drugowitsch and Alwyn M Barry}, title = {A Formal Framework and Extensions for Function Approximation in Learning Classifier Systems}, institution = {University of Bath}, year = 2006, number = {CSBU2006-01}, } @INPROCEEDINGS{duchitert03undemocratic, author = {Wlodzislaw Duch and Lukasz Itert}, title = {Committees of Undemocratic Competent Models}, booktitle = {International Conference on Artificial Neural Networks (ICANN) and International Conference on Neural Information Processing (ICONIP)}, year = 2003, address = {Istanbul, Turkey}, month = June, pages = {33--36}, url = {http://www.phys.uni.torun.pl/publications/kmk/03-C-Ensambles-s.html}, } @BOOK{duda01book, author = {Richard Duda and Peter Hart and David Stork}, title = {Pattern Classification}, publisher = {John Wiley and Sons}, year = 2001, note = {0-471-05669-3}, } @ARTICLE{fern03onlinebranch, author = {Alan Fern and Robert Givan}, title = {Online Ensemble Learning: An Empirical Study.}, journal = {Machine Learning}, volume = 53, number = {1-2}, year = 2003, pages = {71-109}, abstract = {Applied online bagging and boosting to branch prediction}, } @ARTICLE{fern2003rph, title = {{Random projection for high dimensional data clustering: A cluster ensemble approach}}, author = {X. Fern and C. Brodley}, journal = {Proceedings of the Twentieth International Conference on Machine Learning}, pages = {186--193}, year = 2003, } @ARTICLE{fleuret04cmi, title = {{Fast Binary Feature Selection with Conditional Mutual Information}}, author = {Fleuret, F.}, journal = {The Journal of Machine Learning Research}, volume = 5, pages = {1531-1555}, year = 2004, publisher = {MIT Press Cambridge, MA, USA}, } @INPROCEEDINGS{freund01averaging, title = {Why Averaging Classifiers can Protect Against Overfitting}, author = {Yoav Freund and Yishay Mansour and Robert Schapire}, booktitle = {Eighth International Workshop on Artificial Intelligence and Statistics}, year = 2001, } @TECHREPORT{friedman03isle, author = {J.H. Friedman and B. Popescu}, title = {Importance Sampling Learning Ensembles}, institution = {Stanford University}, year = 2003, month = {September}, url = {http://www-stat.stanford.edu/\~{}jhf/ftp/isle.pdf}, } @ARTICLE{gencay01pricing, author = {R. Gencay and Min Qi}, title = {Pricing and hedging derivative securities with neural networks: Bayesian regularization, early stopping, and bagging}, journal = {IEEE Transactions on Neural Networks}, year = 2001, volume = 12, issue = 4, month = {July}, pages = {726--734}, } @INPROCEEDINGS{grandvalet01bagging, author = {Y.~Grandvalet}, title = {Bagging can stabilize without reducing variance}, booktitle = {ICANN'01}, series = {Lecture Notes in Computer Science}, publisher = {Springer}, year = 2001, } @INPROCEEDINGS{navone2000, author = {H.D.Navone and P.F.Verdes and P.M.Granitto and H.A.Ceccatto}, title = {A New Algorithm for Selecting Diverse Members of a Neural Network Ensemble}, booktitle = {6th International Congress on Information Engineering}, year = 2000, note = {Buenos Aires, Argentina}, } @PHDTHESIS{hansen00thesis, author = {J.V. Hansen}, title = {Combining Predictors: Meta Machine Learning Methods and Bias/Variance and Ambiguity Decompositions}, school = {Aarhus Universitet, Datalogisk Institut}, year = 2000, } @ARTICLE{hayashi02medical, author = {Y. Hayashi and R. Setiono}, title = {Combining neural network predictions for medical diagnosis}, journal = {Computers in Biology and Medicine}, year = 2002, volume = 32, number = 4, pages = {237--246}, } @ARTICLE{islam03constructive, author = {Md. M. Islam and X. Yao and K. Murase}, number = 4, year = 2003 , journal = {IEEE Transactions on Neural Networks}, title = {A constructive algorithm for training cooperative neural network ensembles}, month = {July}, pages = {820--834}, volume = 14, } @ARTICLE{itqon01, author = {Itqon and Shun'ichi Kaneko and Satoru Igarashi}, title = {Combining Multiple k-Nearest Neighbor Classifiers Using Feature Combinations}, journal = {Journal of IECI (Indonesian Society on Electrical Electronics, Communication and Information)}, year = 2000, volume = 2, number = 3, pages = {23--31}, } @INPROCEEDINGS{wuzhouchen02:ensembling, author = {J.X.Wu and Z.H.Zhou and Z.Q.Chen}, title = "Ensemble of {GA} based selective neural network, ensembles" , booktitle = {8th International Conference on Neural Information Processing (ICONIP)}, volume = 3, pages = {1477-1482}, year = 2002, } @INBOOK{jaakkola00variational, author = {Jaakkola, T.}, editor = {D. Saad and M. Opper}, title = {Advanced Mean Field methods - Theory and Practice}, chapter = {Tutorial on Variational Approximation Methods}, publisher = {MIT Press}, year = 2000, } @ARTICLE{james03bv, author = {Gareth James}, year = 2003, title = {Variance and Bias for General Loss Functions}, journal = {Machine Learning}, volume = 51, pages = {115--135}, } @INPROCEEDINGS{jin03regularizer, author = {Rong Jin and Yan Liu and Luo Si and Jaime Carbonell and Alexander Hauptmann}, title = {A New Boosting Algorithm using Input-Dependent Regularizer}, booktitle = {20th International Conference on Machine Learning}, year = 2003, } @INPROCEEDINGS{joshi01evaluating, author = {Mahesh V. Joshi and Vipin Kumar and Ramesh C. Agarwal}, title = {Evaluating Boosting Algorithms to Classify Rare Classes: Comparison and Improvements}, booktitle = {{ICDM}}, pages = {257-264}, year = 2001, url = {http://citeseer.nj.nec.com/joshi01evaluating.html}, } @INPROCEEDINGS{kanamori02mixture, author = {Takafumi Kanamori}, title = {A New Sequential Algorithm for Regression Problems by using Mixture Distribution}, booktitle = {Int. Conf. Artif. Neur. Netw. ICANN}, pages = {535--540}, year = 2002, publisher = {Springer}, url = {http://citeseer.nj.nec.com/573689.html}, } @INPROCEEDINGS{khareyao02, author = {Vineet Khare and Xin Yao}, title = {Artificial Speciation of Neural Network Ensembles}, booktitle = {Proc. of the 2002 UK Workshop on Computational Intelligence (UKCI'02)}, pages = {96--103}, year = 2002, editor = {J.A.Bullinaria}, month = {September}, organization = {University of Birmingham, UK}, } @ARTICLE{kleinberg00algorithmic, author = {Eugene M. Kleinberg}, title = {On the Algorithmic Implementation of Stochastic Discrimination}, journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence}, volume = 22, number = 5, pages = {473-490}, year = 2000, } @INPROCEEDINGS{kuncheva00independence, author = {L. Kuncheva and C. Whitaker and C. Shipp and R. Duin}, title = {Is independence good for combining classifiers}, booktitle = {Proceedings of the 15th International Conference on Pattern Recognition, Barcelona, Spain}, year = 2000, pages = {168-171}, } @ARTICLE{kuncheva03limits, author = {L. Kuncheva and C. Whitaker and C. Shipp and R. Duin}, title = {Limits on the majority vote accuracy in classifier fusion}, journal = {Pattern {A}nalysis and {A}pplications}, year = 2003, month = {April}, volume = 6, number = 1, pages = {22--31}, } @ARTICLE{kuncheva02sixstrategies, author = {Kuncheva, L.I.}, title = {A Theoretical Study on Six Classifier Fusion Strategies}, journal = {PAMI}, volume = 24, year = 2002, number = 2, month = {February}, pages = {281-286}, } @ARTICLE{kuncheva02switching, author = {L.I. Kuncheva}, title = {Switching between selection and fusion in combining classifiers: An experiment}, journal = {IEEE Transactions On Systems Man And Cybernetics}, volume = 32, number = 2, pages = {146--156}, year = 2002, } @INPROCEEDINGS{kuncheva03elusive, author = {Kuncheva, L.I.}, title = {That {E}lusive {D}iversity in {C}lassifier {E}nsembles}, booktitle = {First Iberian Conference on Pattern Recognition and Image Analysis (IbPRIA), available as LNCS volume 2652}, pages = {1126--1138}, year = 2003, } @ARTICLE{kuncheva02generating, author = {L.I. Kuncheva and R.K. Kountchev}, title = {Generating Classifier Outputs of Fixed Accuracy and Diversity}, journal = {Pattern Recognition Letters}, year = 2002, number = 23, pages = {593-600}, } @ARTICLE{kunchevawhitaker03, author = {L.I. Kuncheva and C. Whitaker}, title = {Measures of Diversity in Classifier Ensembles}, journal = {Machine Learning}, year = 2003, number = 51, pages = {181--207}, } @ARTICLE{kuncheva2004udc, title = {{Using diversity in cluster ensembles}}, author = {Kuncheva, LI and Hadjitodorov, ST}, journal = {Systems, Man and Cybernetics, 2004 IEEE International Conference on}, volume = 2, year = 2004, } @ARTICLE{hadjitodorov2005mdb, title = {{Moderate Diversity for Better Cluster Ensembles}}, author = {Hadjitodorov, S.T. and Kuncheva, L.I. and Todorova, L.P.}, journal = {Information Fusion}, volume = 7, number = 3, year = 2006, } @INCOLLECTION{kuncheva01:tenmeasures, author = {L.I. Kuncheva and C.J. Whitaker}, title = {Ten Measures of Diversity in Classifier Ensembles: Limits for Two Classifiers}, booktitle = {IEE Workshop on Intelligent Sensor Processing}, publisher = {IEE}, year = 2001, month = {February}, } @BOOK{kunchevabook, author = {Ludmila Kuncheva}, title = {Combining Pattern Classifiers: Methods and Algorithms}, publisher = {Wiley Press}, year = 2004, note = {ISBN 0-471-21078-1}, } @TECHREPORT{kutin01stability, author = {Samuel Kutin and Partha Niyogi}, title = {The Interaction of Stability and Weakness in AdaBoost}, institution = {The University of Chicago}, year = 2001, number = {TR-2001-30}, url = {http://www.cs.uchicago.edu/research/publications/techreports/TR-2001-30}, } @TECHREPORT{langdon:2001:edf, author = {W. B. Langdon}, title = {Evolutionary Data Fusion}, institution = {University College, London}, year = 2001, number = {RN/01/19}, address = {UK}, month = {3 April}, keywords = {genetic algorithms, genetic programming, ROC}, url = {http://www.cs.ucl.ac.uk/staff/W.Langdon/datafusion.html}, url = {http://www.cs.ucl.ac.uk/staff/W.Langdon/roc}, size = {12 pages}, notes = {Distributed at 25 April 2001 Faraday meeting http://www.npl.co.uk/intersect/ }, } @INPROCEEDINGS{langdon:2001:wsc6, author = {W. B. Langdon and S. J. Barrett and B. F. Buxton}, title = {Genetic Programming for Combining Neural Networks for Drug Discovery}, booktitle = {Soft Computing and Industry Recent Applications}, year = 2001, editor = {Rajkumar Roy and Mario K\"oppen and Seppo Ovaska and Takeshi Furuhashi and Frank Hoffmann}, pages = {597--608}, month = {10--24 September}, publisher = {Springer-Verlag}, note = {Published 2002}, keywords = {genetic algorithms, genetic programming, data fusion, data mining, knowledge discovery, Receiver Operating Characteristics, ensemble of classifiers, size fair crossover}, isbn = {1-85233-539-4}, url = {http://www.springer.de/cgi/svcat/search_book.pl?isbn=1-85233-539-4}, } @INPROCEEDINGS{langdon03gp title = {Comparison of AdaBoost and Genetic Programming for combining Neural Networks for Drug Discovery}, author = {W. B. Langdon and S. J. Barrett and B. F. Buxton}, booktitle = {Applications of Evolutionary Computing, EvoWorkshops2003: Evo{BIO}, Evo{COP}, Evo{IASP}, Evo{MUSART}, Evo{ROB}, Evo{STIM}}, editor = {G\"unther R.~Raidl and Stefano Cagnoni and Juan Jes\'us Romero Cardalda and David W.~Corne and Jens Gottlieb and Agn\`es Guillot and Emma Hart and Colin G.~Johnson and Elena Marchiori and Jean-Arcady Meyer and Martin Middendorf}, volume = 2611, series = {LNCS}, pages = {87--98}, address = {University of Essex, UK}, publisher = {Springer-Verlag}, publisher_address ={Berlin}, month = {14-16 April}, organisation = {EvoNet}, year = 2003, keywords = {genetic algorithms, genetic programming, adaboost, drug design, Receiver Operating Characteristics (ROC), ensemble of classifiers, data fusion, artificial neural networks, clementine, high through put screening (HTS)}, size = {12 pages}, notes = {EvoWorkshops2003}, } @INPROCEEDINGS{langdon:2001:gROC, title = {Genetic Programming for Combining Classifiers}, author = {W. B. Langdon and B. F. Buxton}, pages = {66--73}, year = 2001, publisher = {Morgan Kaufmann}, booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001)}, editor = {Lee Spector and Erik D. Goodman and Annie Wu and W.B. Langdon and Hans-Michael Voigt and Mitsuo Gen and Sandip Sen and Marco Dorigo and Shahram Pezeshk and Max H. Garzon and Edmund Burke}, address = {San Francisco, California, USA}, publisher_address ={San Francisco, CA 94104, USA}, month = {7-11 July}, keywords = {genetic algorithms, genetic programming, data fusion, data mining, knowledge discovery, Receiver Operating Characteristics, ensemble of classifiers, size fair crossover}, isbn = {1-55860-774-9}, url = {ftp://cs.ucl.ac.uk/genetic/papers/WBL_gecco2001_roc.ps.gz}, url = {ftp://cs.ucl.ac.uk/genetic/papers/WBL_gecco2001_roc.pdf}, size = {8 pages}, abstract = {Genetic programming (GP) can automatically fuse given classifiers to produce a combined classifier whose Receiver Operating Characteristics (ROC) are better than scott:1998:BMVC ``Maximum Realisable Receiver Operating Characteristics'' (MRROC). I.e. better than their convex hull. This is demonstrated on artificial, medical and satellite image processing bench marks.}, notes = {GECCO-2001 A joint meeting of the tenth International Conference on Genetic Algorithms (ICGA-2001) and the sixth Annual Genetic Programming Conference (GP-2001) Part of spector:2001:GECCO}, } @MISC{langdon:2002:kdmdd, author = {W. B. Langdon and B. F. Buxton and S. J. Barrett}, title = {Combining Machine Learning techniques to Predict Compounds' Cytochrome P450 High Throughput Screening Inhibition}, howpublished = {Knowledge Discovery meets Drug Discovery}, year = 2002, month = {23 October}, note = {poster}, keywords = {genetic algorithms, genetic programming}, url = {ftp://cs.ucl.ac.uk/genetic/papers/wbl_kdmdd2002.pdf}, notes, {http://www.kdnet.org/workshop_overview1_bioinfoLeuven02.htm}, } @INPROCEEDINGS{langdon02gp, year = 2002, title = {A Hybrid Genetic Programming Neural Network Classifier for Use in Drug Discovery}, institution = {Department of Computer Science -- University College London -- UK}, booktitle = {Soft Computing Systems - Design, Management and Applications}, author = {William B. Langdon}, abstract = {We have shown genetic programming (GP) can automatically fuse given classifiers of diverse types to produce a hybrid classifier. Combinations of neural networks, decision trees and Bayes classifier shave been formed. On a range of benchmarks the evolved multiple classifier system is better than all of its components. Indeed its Receiver Operating Characteristics (ROC) are better than [Scott et al., 1998]s "Maximum Realisable Receiver Operating Characteristics" MRROC (convex hull) An important component in the drug discovery is testing potential drugs for activity with P450 cell membrane molecules. Our technique has been used in a blind trial where artificial neural networks are trained by Clementine on P450 pharmaceutical data. Using just the trained networks, GP automatically evolves a composite classifier. Recent experiments with boosting the networks will be compared with genetic programming.}, } @INPROCEEDINGS{langdon:2002:EuroGP, title = {Combining Decision Trees and Neural Networks for Drug Discovery}, author = {William B. Langdon and S. J. Barrett and B. F. Buxton}, booktitle = {Genetic Programming, Proceedings of the 5th European Conference, EuroGP 2002}, pages = {60--70}, address = {Kinsale, Ireland}, publisher_address ={Berlin}, month = {3-5 April}, year = 2002, keywords = {genetic algorithms, genetic programming, drug design, Receiver Operating Characteristics (ROC), ensemble of classifiers, data fusion, artificial neural networks, clementine, decision trees C4.5, high through put screening (HTS)}, isbn = {3-540-43378-3}, size = {10 pages}, abstract = {Genetic programming (GP) offers a generic method of automatically fusing together classifiers using their receiver operating characteristics (ROC) to yield superior ensembles. We combine decision trees (C4.5) and artificial neural networks (ANN) on a difficult pharmaceutical data mining (KDD) drug discovery application. Specifically predicting inhibition of a P450 enzyme. Training data came from high throughput screening (HTS) runs. The evolved model may be used to predict behaviour of virtual (i.e. yet to be manufactured) chemicals. Measures to reduce over fitting are also described. }, notes = {EuroGP'2002, part of lutton:2002:GP}, } @INPROCEEDINGS{langdon:2001:eROC, author = {William B. Langdon and Bernard F. Buxton}, title = {Evolving Receiver Operating Characteristics for Data Fusion}, booktitle = {Genetic Programming, Proceedings of EuroGP'2001}, year = 2001, volume = 2038, pages = {87--96}, address = {Lake Como, Italy}, publisher_address ={Berlin}, month = {18-20 April}, organisation = {EvoNET}, publisher = {Springer-Verlag}, keywords = {genetic algorithms, genetic programming, Data Fusion, Data Mining, Knowledge Discovery, Receiver Operating Characteristics, ROC, Combining Classifiers}, isbn = {3-540-41899-7}, url = {http://evonet.dcs.napier.ac.uk/eurogp2001/}, url = {ftp://cs.ucl.ac.uk/genetic/papers/wbl_egp2001.ps.gz}, size = {10 pages}, abstract = {It has been suggested that the ``Maximum Realisable Receiver Operating Characteristics'' for a combination of classifiers is the convex hull of their individual ROCs [Scott et al., 1998]. As expected in at least some cases better ROCs can be produced. We show genetic programming (GP) can automatically produce a combination of classifiers whose ROC is better than the convex hull of the supplied classifier's ROCs.}, notes = {EuroGP'2001, part of miller:2001:gp}, } @MISC{lappalainen00ensemble, author = {H. Lappalainen and J. Miskin}, title = {Ensemble Learning}, text = {H. Lappalainen and J. Miskin, Ensemble Learning, in M. Girolami (Ed.), Advances in Independent Component Analysis, Springer, Berlin, 2000 (in press).}, year = 2000, } @ARTICLE{lee2004lob, title = {{Lossless Online Bayesian Bagging}}, author = {Lee, H.K.H. and Clyde, M.A.}, journal = {Journal of Machine Learning Research}, volume = 5, pages = {143-151}, year = 2004, publisher = {MIT Press Cambridge, MA, USA}, } @ARTICLE{liu00evolutionary, author = {Y. Liu and X. Yao and T. Higuchi}, title = {Evolutionary Ensembles with Negative Correlation Learning}, journal = {IEEE Transactions on Evolutionary Computation}, volume = 4, number = 4, month = {November}, year = 2000, url = {http://citeseer.nj.nec.com/article/liu00evolutionary.html}, } @INPROCEEDINGS{liuyao02decision, author = {Y. Liu and X. Yao and Q. Zhao and T. Higuchi}, title = {An experimental comparison of neural network ensemble learning methods on decision boundaries}, booktitle = {Proceedings of the 2002 International Joint Conference on Neural Networks (IJCNN'02)}, pages = {221-226}, month = {May}, year = 2002, publisher = {IEEE Press, Piscataway, NJ, USA}, } @INPROCEEDINGS{liu01mutual, author = {Yong Liu and Xin Yao and Qiangfu Zhao and Tetsuya Higuchi}, title = {Evolving a Cooperative Population of Neural Networks by Minimizing Mutual Information}, booktitle = {Proceedings of the 2001 Congress on Evolutionary Computation}, pages = {384-389}, year = 2001, month = {May}, publisher = {IEEE Press}, } @INPROCEEDINGS{luochen02b, author = {Dingsheng Luo and Ke Chen}, title = {On the use of statistical ensemble methods for telephone-line speaker identification}, booktitle = {Proceedings of International Joint Conference on Communications, Circuits and Systems (ICCCAS'2002)}, year = 2002, publisher = {IEEE Press}, address = {Chengdu, China}, pages = {II904-II908}, month = {July}, } @INPROCEEDINGS{luochen02a, author = {Dingsheng Luo and Ke Chen}, title = {A comparative study of statistical ensemble methods on mismatch conditions}, booktitle = {Proceedings of World Congress on Computational Intelligence: International Joint Conference on Neural Networks (WCCI 2002 and IJCNN 2002)}, year = 2002, address = {Honolulu USA}, publisher = {IEEE Press}, pages = {59--64}, month = {November}, } @ARTICLE{malzahn03approximate, author = {Dorthe Malzahn and Manfred Opper}, title = {An Approximate Analytical Approach to Resampling Averages}, journal = {Journal of Machine Learning Research}, year = 2003, volume = 4, pages = {1151--1173}, month = {December}, url = {http://www.jmlr.org}, } @INBOOK{mason00functional, author = {L. Mason and J. Baxter and P. L. Bartlett and M. Frean}, editor = {A.J. Smola, P. L. Bartlett, B. Scholkopf, and D. Schuurmans}, title = {Advances in Large Margin Classifiers : Functional gradient techniques for combining hypotheses}, publisher = {MIT Press}, year = 2000, address = {Cambridge, MA}, pages = {221--246}, } @INPROCEEDINGS{mckay00sharing, author = {Bob McKay}, title = {Fitness Sharing in Genetic Programming}, pages = {435--442}, year = 2000, publisher = {Morgan Kaufmann}, booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2000)}, editor = {Darrell Whitley and David Goldberg and Erick Cantu-Paz and Lee Spector and Ian Parmee and Hans-Georg Beyer}, address = {Las Vegas, Nevada, USA}, publisher_address ="San Francisco, CA 94104, USA", month = "10-12 " # jul, keywords = {genetic algorithms, genetic programming}, isbn = {1-55860-708-0}, notes = {A joint meeting of the ninth International Conference on Genetic Algorithms (ICGA-2000) and the fifth Annual Genetic Programming Conference (GP-2000) Part of whitley:2000:GECCO}, } @INPROCEEDINGS{mckayabbass01analyzing, author = {R. McKay and H. Abbass}, title = {Analyzing Anticorrelation in Ensemble Learning}, booktitle = {Proceedings of 2001 Conference on Artificial Neural Networks and Expert Systems}, year = 2001, pages = {22--27}, address = {Otago, New Zealand}, } @INPROCEEDINGS{mckayabbass01rtqrt, author = {Robert McKay and Hussein Abbass}, title = {Anticorrelation Measures in Genetic Programming}, booktitle = {Australasia-Japan Workshop on Intelligent and Evolutionary Systems}, pages = {45--51}, year = 2001, } @INPROCEEDINGS{meir00localized, author = {Ron Meir and Ran El-{Y}aniv and Shai Ben-{D}avid}, title = {Localized Boosting}, booktitle = {Proc. 13th Annu. Conference on Comput. Learning Theory}, publisher = {Morgan Kaufmann, San Francisco}, pages = {190--199}, year = 2000, url = {http://citeseer.nj.nec.com/511402.html}, } @article{melville05artificial, author = {Prem Melville and Raymond J. Mooney}, title = {Creating diversity in ensembles using artificial data.}, journal = {Information Fusion}, volume = {6}, number = {1}, year = {2005}, pages = {99-111} } @phdthesis{melville05thesis, author = {Prem Melville}, title = {Creating Ensemble Diversity to Reduce Supervision}, institution = {University of Texas at Austin} year = {2005} } @INPROCEEDINGS{melville03decorate, author = {Prem Melville and Ray Mooney}, title = {Constructing Diverse Classifier Ensembles Using Artificial Training Examples}, booktitle = {Proceedings of the Eighteenth International Joint Conference on Artificial Intelligence}, pages = {505--510}, year = 2003, address = {Mexico}, month = {August}, } @article{galor05assessing, author = {Mordechai Gal-Or and Jerrold H. May and William E. Spangler}, title = {Assessing the predictive accuracy of diversity measures with domain-dependent, asymmetric misclassification costs.}, journal = {Information Fusion}, volume = {6}, number = {1}, year = {2005}, pages = {37-48} } @ARTICLE{rooney2006pes, title = {{Pruning extensions to stacking}}, author = {Rooney, N. and Patterson, D. and Nugent, C.}, journal = {Intelligent Data Analysis}, volume = 10, number = 1, pages = {47-66}, year = 2006, publisher = {IOS Press}, abstract = {In this paper we investigate an algorithmic extension to the technique of Stacking for regression that prunes the ensemble set before application based on a consideration of the training accuracy and diversity of the ensemble members. We evaluate two variants of this approach in comparison to the standard Stacking algorithm, one of which is a static approach that prunes back the ensemble to the same constant size; the other of which is a variable approach prunes the ensemble to an appropriate level based on measures of accuracy and diversity of the ensemble members. We show that on average both techniques are robust in performance to their non-pruned counterpart, while having the advantage of producing smaller and less complex ensembles. In the latter respect, the static approach proved more effective, but we show that the variable approach lends itself better for further optimization.}, } @ARTICLE{dettling2003btc, title = {{Boosting for tumor classification with gene expression data}}, author = {Dettling, M. and B{\"u}hlmann, P.}, journal = {Bioinformatics}, volume = 19, number = 9, pages = {1061-1069}, year = 2003, } @ARTICLE{eibl06multi, title = {{Multiclass Boosting for Weak Classifiers}}, author = {Eibl, G. and Pfeiffer, K.P.}, journal = {The Journal of Machine Learning Research}, volume = 6, pages = {189-210}, year = 2005, publisher = {MIT Press Cambridge, MA, USA}, } @ARTICLE{friedman2000sip, title = {{Special Invited Paper. Additive Logistic Regression: A Statistical View of Boosting}}, author = {Friedman, J. and Hastie, T. and Tibshirani, R.}, journal = {The Annals of Statistics}, volume = 28, number = 2, pages = {337-374}, year = 2000, publisher = {JSTOR}, } @ARTICLE{torralba2004sfe, title = {{Sharing features: efficient boosting procedures for multiclass object detection}}, author = {Torralba, A. and Murphy, KP and Freeman, WT}, journal = {Computer Vision and Pattern Recognition, 2004}, volume = 2, } @ARTICLE{folino2006icg, title = {{Improving cooperative GP ensemble with clustering and pruning for pattern classification}}, author = {Folino, G. and Pizzuti, C. and Spezzano, G.}, journal = {Proceedings of the 8th annual conference on Genetic and evolutionary computation}, pages = {791-798}, year = 2006, publisher = {ACM Press New York, NY, USA}, } @ARTICLE{bi2006eae, title = {{An evidential approach in ensembles}}, author = {Bi, Y. and Dubitzky, W.}, journal = {Proceedings of the 2006 ACM symposium on Applied computing}, pages = {1-6}, year = 2006, publisher = {ACM Press New York, NY, USA}, abstract = {In this paper, we describe an approach to modeling the general process of combining decisions involved in ensembles of classifiers as an evidential reasoning process. This work proposes a novel structure, theoretical properties and manipulation mechanisms for representing classifier decisions as pieces of evidence. The advantage of the representation formalism is that it not only facilitates the distinguishing of trivial focal elements from important ones, resulting in the improvement of the ensemble performance, but it also effectively reduces the computation-time from exponential (as required in the conventional process of combining multiple pieces of evidence) to linear. We have conducted a comparative analysis on the effectiveness of the proposed evidence representation formalism in the text categorization domain. By comparing this method with majority voting and the previous results, we also demonstrate the advantage of this novel approach in combining classifiers.}, } @ARTICLE{mashao2006ccd, title = {{Combining Classifier Decisions for Robust Speaker Identification}}, author = {MASHAO, D.J. and SKOSAN, M.}, journal = {Pattern recognition}, volume = 39, number = 1, pages = {147-155}, year = 2006, publisher = {Elsevier Science}, } @ARTICLE{murua02upperbound, author = {Murua, A.}, title = {Upper Bounds for Error Rates of Linear Combinations of Classifiers}, journal = {PAMI}, volume = 24, year = 2002, number = 5, month = {May}, pages = {591-602}, } @ARTICLE{nanni2006ekl, title = {{An ensemble of K-local hyperplanes for predicting protein-protein interactions}}, author = {Nanni, L. and Lumini, A.}, journal = {Bioinformatics}, volume = 22, number = 10, pages = 1207, year = 2006, } @ARTICLE{nair:jmlr02, title = {Some greedy learning algorithms for sparse regression and classification with mercer kernels}, author = {P. B. Nair and A. Choudhury and A. J. Keane}, journal = {Journal of Machine Learning Research}, volume = 3, pages = {781-801}, year = 2002, } @PHDTHESIS{oza2001thesis, title = {{Online Ensemble Learning}}, author = {Oza, N.C.}, year = 2001, school = {UNIVERSITY of CALIFORNIA}, } @ARTICLE{oza2001eco, title = {{Experimental comparisons of online and batch versions of bagging and boosting}}, author = {Oza, N.C. and Russell, S.}, journal = {Proceedings of the seventh ACM SIGKDD international conference on Knowledge discovery and data mining}, pages = {359-364}, year = 2001, publisher = {ACM Press New York, NY, USA}, } @ARTICLE{oza2005oba, title = {{Online bagging and boosting}}, author = {Oza, NC}, journal = {Systems, Man and Cybernetics, 2005 IEEE International Conference on}, volume = 3, year = 2005, } @INPROCEEDINGS{pennock00normative, author = {David M. Pennock and Pedrito {Maynard-Reid {II}} and C. Lee Giles and Eric Horvitz}, title = {A Normative Examination of Ensemble Learning Algorithms}, booktitle = {Proc. 17th International Conf. on Machine Learning}, publisher = {Morgan Kaufmann, San Francisco, CA}, pages = {735--742}, year = 2000, } @TECHREPORT{poggio02bagging, author = {Tomaso Poggio and Ryan Rifkin and Sayan Mukherjee and Alex Rakhlin}, title = {Bagging Regularizes}, institution = {MIT AI Lab}, year = 2002, number = {AI Memo 2002-003, CBCL Memo 214}, } @MISC{roli02notes, author = {Fabio Roli}, title = {Lecture Notes: Linear Combiners for Fusion of Pattern Classifiers}, url = {http://www.disi.unige.it/person/MasulliF/ricerca/school2002/contributions/vietri02-lect-roli.pdf}, } @article{ruta05selection, author = {Dymitr Ruta and Bogdan Gabrys}, title = {Classifier selection for majority voting.}, journal = {Information Fusion}, volume = {6}, number = {1}, year = {2005}, pages = {63-81} } @ARTICLE{ruta03sets, author = {Dymitr Ruta and Bogdan Gabrys}, title = {Set Analysis of Coincident Errors and Its Applications for Combining Classifiers}, journal = {Combinatorial Optimisation}, volume = 13, publisher = {Kluwer Academic}, year = 2003, } @ARTICLE{schapiresinger00boostexter, author = {R.E.~Schapire and Y.~Singer}, title = {BoosTexter: A boosting-based system for text categorization}, journal = {Machine Learning}, year = 2000, volume = 39, number = {2/3}, pages = {135-168}, } @INPROCEEDINGS{shan05cmi, author = {Shan, C. and Gong, S. and McOwan, P.W.}, title = {{Conditional Mutual Information Based Boosting for Facial Expression Recognition}}, booktitle = {BMVC}, year = 2005 } @INPROCEEDINGS{shih03bundling, author = {Lawrence Shih and Jason Rennie and Yu-Han Chang and David Karger}, title = {Text Bundling : Statistics Based Data Reduction}, booktitle = {20th {I}nternational {C}onference on {M}achine {L}earning (ICML'03)}, year = 2003, editor = {Tom Fawcett and Nina Mishra}, month = {August}, address = {Washington DC, USA}, } @ARTICLE{shippkuncheva02, author = {C.A. Shipp and L.I. Kuncheva}, title = {Relationships between combination methods and measures of diversity in combining classifiers}, journal = {Information Fusion}, year = 2002, number = 3, pages = {135-148}, } @INPROCEEDINGS{skurichina00role, author = {M. Skurichina and R.P.W. Duin}, title = {The Role of Combining Rules in Bagging and Boosting}, booktitle = {Advances in Pattern Recognition, Proc. Joint IAPR International Workshops SSPR2000 and SPR2000 Lecture Notes in Computer Science, vol. 1876}, publisher = {Springer}, month = {June}, pages = {236-245}, editors = {F.J. Ferri, J.M. Inesta, A. Amin, P. Pudil}, year = 2000, address = {Alicante, Spain}, } @PHDTHESIS{schubert2005qca, title = {{Quantifying correlation and its effects on system performance in classifier fusion}}, author = {Schubert, C.M.}, year = 2005, school = {AIR FORCE INSTITUTE OF TECHNOLOGY}, url = {http://gradworks.umi.com/31/91/3191355.html} } @ARTICLE{skurichina2002bba, title = {{Bagging, Boosting and the Random Subspace Method for Linear Classifiers}}, author = {Skurichina, M.W. and Duin, R.P.W.W.}, journal = {Pattern Analysis \& Applications}, volume = 5, number = 2, pages = {121-135}, year = 2002, publisher = {Springer}, } @INPROCEEDINGS{stainvas00, title = {Blurred Face Recognition via a Hybrid Network Architecture}, author = {I. Stainvas and N. Intrator}, booktitle = {ICPR}, year = 2000, volume = 2, pages = {809--812}, } @INPROCEEDINGS{stephenson03compiler, author = {Mark Stephenson and Una-May O'Reilly and Martin C. Martin and Saman P. Amarasinghe}, title = {Genetic Programming Applied to Compiler Heuristic Optimization.}, booktitle = {EuroGP}, year = 2003, pages = {238-253}, } @ARTICLE{strehl2003cek, title = {{Cluster ensembles- A knowledge reuse framework for combining multiple partitions.}}, author = {Strehl, A. and Ghosh, J.}, journal = {Journal of Machine Learning Research}, volume = 3, number = 3, pages = {583-617}, year = 2003, publisher = {MIT Press}, } @ARTICLE{suganthan01hierarchical, author = {P. N. Suganthan}, title = {Pattern classification using multiple hierarchical overlapped self-organising maps}, journal = {Pattern Recognition}, year = 2001, volume = 34, number = 11, pages = {2173-2179}, month = {November}, } @ARTICLE{tetko02associative, author = {Tetko, I. V.}, title = {Associative neural network}, journal = {Neural Processing Letters}, volume = 16, number = 2, pages = {187-199}, abstract = {An associative neural network (ASNN) is a combination of an ensemble of the feed-forward neural networks and the K-nearest neighbor technique. The introduced network uses correlation between ensemble responses as a measure of distance among the analyzed cases for the nearest neighbor technique and provides an improved prediction by the bias correction of the neural network ensemble both for function approximation and classification. Actually, the proposed method corrects a bias of a global model for a considered data case by analyzing the biases of its nearest neighbors determined in the space of calculated models. An associative neural network has a memory that can coincide with the training set. If new data become available the network can provide a reasonable approximation of such data without a need to retrain the neural network ensemble. Applications of ASNN for prediction of lipophilicity of chemical compounds and classification of UCI letter and satellite data set are presented. The developed algorithm is available on-line at http://www.virtuallaboratory.org/lab/asnn.}, year = 2002, url = {http://cogprints.ecs.soton.ac.uk/archive/00001441/}, } @ARTICLE{tetko02associative2, author = {Tetko, I. V.}, title = {Neural network studies. 4. Introduction to associative neural networks}, journal = {J Chem Inf Comput Sci}, volume = 42, number = 3, pages = {717-28.}, keywords = {*Neural Networks (Computer) Support, Non-U.S. Gov\'t}, abstract = {Associative neural network (ASNN) represents a combination of an ensemble of feed-forward neural networks and the k-nearest neighbor technique. This method uses the correlation between ensemble response,, as a measure of distance amid the analyzed cases for the nearest neighbor technique. This provides an improved prediction by the bias correction of the neural network ensemble. An associative neural network has a memory that can coincide with the training set. If new data becomes available, the network further improves its predictive ability and provides a reasonable approximation of the unknown function without a need to retrain the neural network ensemble, This feature of the method dramatically improves its predictive ability over traditional neural networks and k-nearest neighbor techniques, as demonstrated using several artificial data sets and a program to predict lipophilicity of chemical compounds. Another important feature of ASNN is the possibility to interpret neural network results by analysis of correlations between data cases in the space of models, It is shown that analysis of such correlations makes it possible to provide \"property-targeted\" clustering of data. The possible applications and importance of ASNN in drug design and medicinal and combinatorial chemistry are discussed. The method is available on-line at http://www.vcclab.org/lab/asnn.}, year = 2002, } @ARTICLE{tetko01volume, author = {Tetko, I. V. and Kovalishyn, V. V. and Livingstone, D. J.}, title = {Volume learning algorithm artificial neural networks for 3D QSAR studies}, journal = {J Med Chem}, volume = 44, number = 15, pages = {2411-20.}, abstract = {The current study introduces a new method, the volume learning algorithm (VLA), for the investigation of three-dimensional quantitative structure-activity relationships (QSAR) of chemical compounds. This method incorporates the advantages of comparative molecular field analysis (CoMFA) and artificial neural network approaches. VLA is a combination of supervised and unsupervised neural networks applied to solve the same problem. The supervised algorithm is a feed-forward neural network trained with a back-propagation algorithm while the unsupervised network is a self-organizing map of Kohonen. The use of both of these algorithms makes it possible to cluster the input CoMFA field variables and to use only a small number of the most relevant parameters to correlate spatial properties of the molecules with their activity. The statistical coefficients calculated by the proposed algorithm for cannabimimetic aminoalkyl indoles were comparable to, or improved, in comparison to the original study using the partial least squares algorithm. The results of the algorithm can be visualized and easily interpreted. Thus, VLA is a new convenient tool for three-dimensional QSAR studies.}, year = 2001, } @ARTICLE{tetko02asnnapp, author = {Tetko, I. V. and Tanchuk, V. Y.}, title = {Application of associative neural networks for prediction of lipophilicity in ALOGPS 2.1 program}, journal = {J Chem Inf Comput Sci}, volume = 42, number = 5, pages = {1136-45.}, abstract = {This article provides a systematic study of several important parameters of the Associative Neural Network (ASNN), such as the number of networks in the ensemble, distance measures, neighbor functions, selection of smoothing parameters, and strategies for the user-training feature of the algorithm. The performance of the different methods is assessed with several training/test sets used to predict lipophilicity of chemical compounds. The Spearman rank-order correlation coefficient and Parzen-window regression methods provide the best performance of the algorithm. If additional user data is available, an improved prediction of lipophilicity of chemicals up to 2-5 times can be calculated when the appropriate smoothing parameters for the neural network are selected. The detected best combinations of parameters and strategies are implemented in the ALOGPS 2.1 program that is publicly available at http://www.vcclab.org/lab/alogps.}, year = 2002, } @ARTICLE{tetko01estimation, author = {Tetko, I. V. and Tanchuk, V. Y. and Kasheva, T. N. and Villa, A. E.}, title = {Estimation of aqueous solubility of chemical compounds using E-state indices}, journal = {J Chem Inf Comput Sci}, volume = 41, number = 6, pages = {1488-93.}, abstract = {The molecular weight and electrotopological E-state indices were used to estimate by Artificial Neural Networks aqueous solubility for a diverse set of 1291 organic compounds. The neural network with 33-4-1 neurons provided highly predictive results with r(2) = 0.91 and RMS = 0.62. The used parameters included several combinations of E-state indices with similar properties. The calculated results were similar to those published for these data by Huuskonen (2000). However, in the current study only E-state indices were used without need of additional indices (the molecular connectivity, shape, flexibility and indicator indices) also considered in the previous study. In addition, the present neural network contained three times less hidden neurons. Smaller neural networks and use of one homogeneous set of parameters provides a more robust model for prediction of aqueous solubility of chemical compounds. Limitations of the developed method for prediction of large compounds are discussed, The developed approach is available online at http://www.vcclab.org/lab/alogps}, year = 2001, } @ARTICLE{tetko01prediction, author = {Tetko, I. V. and Tanchuk, V. Y. and Villa, A. E.}, title = {Prediction of n-octanol/water partition coefficients from PHYSPROP database using artificial neural networks and E-state indices}, journal = {J Chem Inf Comput Sci}, volume = 41, number = 5, pages = {1407-21.}, abstract = {The molecular weight and electrotopological E-state indices were used to estimate by Artificial Neural Networks aqueous solubility for a diverse set of 1291 organic compounds. The neural network with 33-4-1 neurons provided highly predictive results with r(2) = 0.91 and RMS = 0.62. The used parameters included several combinations of E-state indices with similar properties. The calculated results were similar to those published for these data by Huuskonen (2000). However, in the current study only E-state indices were used without need of additional indices (the molecular connectivity, shape, flexibility and indicator indices) also considered in the previous study. In addition, the present neural network contained three times less hidden neurons. Smaller neural networks and use of one homogeneous set of parameters provides a more robust model for prediction of aqueous solubility of chemical compounds. Limitations of the developed method for prediction of large compounds are discussed. The developed approach is available online at http://www.vcclab.org/lab/alogps.}, year = 2001, } @article{tsymbal05search, author = {Alexey Tsymbal and Mykola Pechenizkiy and Padraig Cunningham}, title = {Diversity in search strategies for ensemble feature selection.}, journal = {Information Fusion}, volume = {6}, number = {1}, year = {2005}, pages = {83-98} } @ARTICLE{tino04nonlinear, author = {P. Tino and I. Nabney and B.S. Williams and J. Losel and Y. Sun}, title = {Non-linear Prediction of Quantitative Structure-Activity Relationships}, journal = {Journal of Chemical Information and Computer Sciences}, year = 2004, volume = 44, number = 5, pages = {1647--1653}, } @ARTICLE{topchy2004acp, title = {{Analysis of consensus partition in cluster ensemble}}, author = {Topchy, AP and Law, MHC and Jain, AK and Fred, AL}, journal = {Data Mining, 2004. ICDM 2004. Proceedings. Fourth IEEE International Conference on}, pages = {225-232}, year = 2004, } @INPROCEEDINGS{tresp00generalizedbayescommittee, author = {Volker Tresp}, title = {The Generalized Bayesian Committee Machine}, booktitle = {Proceedings of the Sixth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD-2000}, year = 2000, pages = {130-139}, abstract = { In this paper we introduce the Generalized Bayesian Committee Machine (GBCM) for applications with large data sets. In particular, the GBCM can be used in the context of kernel based systems such as smoothing splines, kriging, regularization networks and Gaussian process regression which ---for computational reasons--- are otherwise limited to rather small data sets. The GBCM provides a novel and principled way of combining estimators trained for regression, classification, the prediction of counts, the prediction of lifetimes and other applications which can be derived from the exponential family of distributions. We describe an online version of the GBCM which only requires one pass through the data set and only requires the storage of a matrix of the dimension of the number of query or test points. After training, the prediction at additional test points only requires resources dependent on the number of query points but is independent of the number of training data. We confirm the good scaling behavior using real and experimental data sets. }, url = {http://www.boosting.org/papers/upload_7240_kddpaper2.ps}, } @ARTICLE{tresp00bayescommittee, author = {Volker Tresp}, title = {A Bayesian Committee Machine}, journal = {Neural Computation}, year = 2000, pages = {2719-2741}, volume = 12, number = 11, abstract = { The Bayesian committee machine (BCM) is a novel approach to combining estimators which were trained on different data sets. Although the BCM can be applied to the combination of any kind of estimators the main foci are Gaussian process regression and related systems such as regularization networks and smoothing splines for which the degrees of freedom increase with the number of training data. Somewhat surprisingly, we find that the performance of the BCM improves if several test points are queried at the same time and is optimal if the number of test points is at least as large as the degrees of freedom of the estimator. The BCM also provides a new solution for online learning with potential applications to data mining. We apply the BCM to systems with fixed basis functions and discuss its relationship to Gaussian process regression. Finally, we also show how the ideas behind the BCM can be applied in a non-Bayesian setting to extend the input dependent combination of estimators.}, url = {http://www.boosting.org/papers/upload_7235_bcm5.ps}, } @INBOOK{tresp01committeemachines, author = {Volker Tresp}, editor = {Yu Hen Hu and Jenq-Nen Hwang}, title = {Handbook for Neural Network Signal Processing}, chapter = {Committee machines}, publisher = {CRC Press}, year = 2001, } @ARTICLE{tumerghosh00robust, author = {K Tumer and J Ghosh}, title = {Robust Combining of Disparate Classifiers through Order Statistics}, journal = {To appear in Pattern Analysis and Applications, Special issue on Fusion of Multiple Classifiers}, year = 2002, number = 2, volume = 5, url = {http://citeseer.nj.nec.com/592387.html}, } @ARTICLE{ueda00optimal, author = {N Ueda}, title = {Optimal linear combination of neural networks for improving classification Performance }, journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence}, year = 2000, volume = 22, number = 2, pages = {207--215}, } @INPROCEEDINGS{valentini03baggedsvm, author = {Giorgio Valentini and Thomas G. Dietterich}, title = {Low Bias Bagged Support Vector Machines}, booktitle = {20th {I}nternational {C}onference on {M}achine {L}earning (ICML'03)}, year = 2003, editor = {Tom Fawcett and Nina Mishra}, month = {August}, address = {Washington DC, USA}, } @ARTICLE{valev01multi, author = {V. Valev and A. Asaithambi}, title = {Multidimensional pattern recognition problems and combining classifiers}, journal = {Pattern Recognition Letters}, volume = 22, year = 2001, number = 12, month = {October}, pages = {1291-1297}, } @ARTICLE{viola01boosting, title = {{Rapid object detection using a boosted cascade of simple features}}, author = {Viola, P. and Jones, M.}, journal = {Proc. CVPR}, volume = 1, pages = {511-518}, year = 2001, } @article{wang2003mcd, title = {{Mining concept-drifting data streams using ensemble classifiers}}, author = {Wang, H. and Fan, W. and Yu, P.S. and Han, J.}, journal = {Proceedings of the ninth ACM SIGKDD international conference on Knowledge discovery and data mining}, pages = {226-235}, year = {2003}, publisher = {ACM Press New York, NY, USA} } @INPROCEEDINGS{wang01diversity, author = {Wenjia Wang and Derek Partridge and John Etherington}, title = {Hybrid Ensembles and Coincident-Failure Diversity}, booktitle = {Proceedings of the International Joint Conference on Neural Networks}, publisher = {IEEE Press}, month = {July}, pages = {2376--2381}, volume = 4, year = 2001, address = {Washington, USA}, } @ARTICLE{webb05nsn, title = {{Not So Naive Bayes: Aggregating One-Dependence Estimators}}, author = {Webb, G.I. and Boughton, J.R. and Wang, Z.}, journal = {Machine Learning}, volume = 58, number = 1, pages = {5-24}, year = 2005, publisher = {Springer}, annote = {Combining probabilistic models - similar to naive bayes.}, } @ARTICLE{webb00multiboosting, author = {Geoffrey I. Webb}, title = {Multi{B}oosting: {A} Technique for Combining {B}oosting and {W}agging}, journal = {Machine Learning}, volume = 40, number = 2, publisher = {Kluwer Academic Publishers, Boston}, pages = {159--196}, year = 2000, url = {http://citeseer.nj.nec.com/webb98multiboosting.html}, } @INPROCEEDINGS{wersing02a, author = {H. Wersing and E. K{\"o}rner}, title = {Unsupervised Learning of Combination Features for Hierarchical Recognition Models}, booktitle = {Int. Conf. Artif. Neur. Netw. ICANN}, year = 2002, note = {accepted}, } @ARTICLE{wersingkorner02, author = {Heiko Wersing and Edgar Korner}, title = {Learning Optimized Features for Hierarchical Models of Invariant Object Recognition}, journal = {Neural Computation (to appear)}, year = 2002, } @INPROCEEDINGS{wezel00nonconformist, author = {M.C. van Wezel and M.D. Out and W.A. Kosters}, title = {Ensembles of nonconformist neural networks}, booktitle = {Proceedings of the Twelfth Belgium-Netherlands Artificial Intelligence Conference}, editors = {H. Weigand and A. van den Bosch}, pages = {165-172}, year = 2000, } @TECHREPORT{whitaker03examining, author = {Christopher Whitaker and Ludmila Kuncheva}, title = {Examining the relationship between majority vote accuracy and diversity in bagging and boosting}, institution = {School of Informatics, University of Wales, Bangor}, year = 2003, type = {Technical Report}, url = {http://www.informatics.bangor.ac.uk/~kuncheva/papers/lkcw_tr.pdf}, } @article{windeatt05measures, author = {Terry Windeatt}, title = {Diversity measures for multiple classifier system analysis and design.}, journal = {Information Fusion}, volume = {6}, number = {1}, year = {2005}, pages = {21-36} } @INPROCEEDINGS{yaobrown01telecoms, author = {X. Yao and M. Fischer and G. Brown}, title = {Neural Network Ensembles and their Application to Traffic Flow Prediction in Telecommunications Networks}, booktitle = {Proceedings of International Joint Conference on Neural Networks}, note = {Washington DC}, pages = {693-698}, publisher = {IEEE Press}, year = 2001, } @INPROCEEDINGS{yang06select, author = {Ying Yang and Geoff Webb and Jesus Cerquides and Kevin Korb and Janice Boughton and Kai Ming Ting}, title = {To Select or To Weigh: A Comparative Study of Model Selection and Model Weighing for SPODE Ensembles}, booktitle = {17th European Conference on Machine Learning (ECML)}, year = 2006, } @INPROCEEDINGS{yang05ensemble, author = {Ying Yang and Kevin Korb and Kai Ming Ting and Geoff Webb}, title = {Ensemble Selection for SuperParent-One-Dependence Estimators}, booktitle = {18th Australian Joint Conference on Artificial Intelligence}, year = 2005, } @ARTICLE{zhouwutang02ensembling, author = {Z.-H. Zhou, J. Wu, and W. Tang}, title = {Ensembling neural networks: Many could be better than all}, journal = {Artificial Intelligence}, volume = 137, number = {1-2}, pages = {239-263}, year = 2002, } @ARTICLE{zhou02extracting, author = {Z.-H. Zhou, Y. Jiang, and S.-F. Chen}, title = {Extracting Symbolic Rules from Trained Neural Network Ensembles}, journal = {AI Communications, 2003, 16(1): 3-15}, volume = 16, number = 1, pages = {3-15}, year = 2003, url = {http://citeseer.nj.nec.com/zhou03extracting.html}, } @ARTICLE{zenobi01using, author = {Gabriele Zenobi and P{\'a}draig Cunningham}, title = {Using Diversity in Preparing Ensembles of Classifiers Based on Different Feature Subsets to Minimize Generalization Error}, journal = {Lecture Notes in Computer Science}, volume = 2167, pages = {576--587}, year = 2001, } @ARTICLE{zhang2004onb, title = {{The optimality of naive Bayes}}, author = {Zhang, H.}, journal = {17th International FLAIRS conference, Miami Beach, May}, pages = {17-19}, year = 2004, } @INPROCEEDINGS{zhou02ensembling, author = {Z.H. Zhou and J.Wu and W.Tang and Z.Q. Chen}, title = "Selectively ensembling neural classifiers" , booktitle = {International Joint Conference on Neural Networks}, volume = 2, page = {1411-1415}, year = 2002, } @ARTICLE{zhu2006ecn, title = {{Effective classification of noisy data streams with attribute-oriented dynamic classifier selection}}, author = {Zhu, X. and Wu, X. and Yang, Y.}, journal = {Knowledge and Information Systems}, volume = 9, number = 3, pages = {339-363}, year = 2006, publisher = {Springer}, } #PRE2000ENSEMBLEPAPERS @TECHREPORT{ali95comparison, author = {K. Ali}, year = 1995, title = {A comparison of methods for learning and combining evidence from multiple models}, institution = {University of California, Irvine, Dept. of Information and Computer Sciences}, number = {UCI TR \#95-47}, abstract = {Most previous work on multiple models has been done on a few domains. We present a comparsion of three ways of learning multiple models on 29 data sets from the UCI repository. The methods are bagging, $k$-fold partition learning and stochastic search. By using 29 data sets of various kinds - artificial data sets, artificial data sets with noise, molecular-biology and real-world noisy data sets - we are able to draw robust experimental conclusions about the kinds of data sets for which each learning method works best. We also compare four evidence combination methods (Uniform Voting, Bayesian Combination, Distribution Summation and Likelihood Combination) and characterize the kinds of data sets for which each method works best.}, } @ARTICLE{avnimelech99boostedregression, author = {R.~Avnimelech and N.~Intrator}, title = {Boosting Regression Estimators}, journal = {Neural Computation}, year = 1999, volume = 11, pages = {491--513}, } @ARTICLE{avnimelechintrator99boosted, pages = {475-490}, author = {Ran Avnimelech and Nathan Intrator}, year = 1999, journal = {Neural Computation}, volume = 11, title = {Boosted Mixtures of Experts: An Ensemble Learning Scheme}, } @ARTICLE{parmanto96, author = {B.Parmanto and P.W.Munro and H.R.Doyle}, title = {Improving Committee Diagnosis with Resampling Techniques}, editor = {D.S.Touretzky and M.C.Mozer and M.E.Hesselmo}, volume = 8, journal = {Advances in Neural Information Processing Systems}, year = 1996, pages = {882--888}, publisher = {The {MIT} Press}, } @ARTICLE{back93evolution, author = {T. B{\"{a}}ck and H.-P. Schwefel}, title = {An overview of evolutionary algorithms for parameter optimization}, journal = {Evolutionary Computation}, year = 1993, volume = 1, number = 1, pages = {1-23}, } @ARTICLE{batesgranger69, author = {J. M. Bates and C. W. J. Granger}, title = {The combination of forecasts}, journal = {Operations Research Quarterly}, year = 1969, number = 20, pages = {451-468}, annote = {The use of combination techniques in financial forecasting.} } @ARTICLE{battiti1994democracy, author = {Roberto Battiti and Anna Maria Colla}, title = {Democracy in Neural Nets: Voting Schemes for Classification}, journal = {Neural Networks}, year = 1994, volume = 7, number = 4, pages = {691--707}, abstract = {Discusses some possible ways to combine the outputs of a set of neural network classifiers to reach a combined decision with a higher performance in terms of lower rejection rates and/or better accuracy rates. The methods considered range from the requirement of a complete agreement among the individual classifications to election schemes based on the distribution of votes collected by the different classes. In addition, the rejection rules based on the different output classes can be complemented by rules that also consider the information in the individual output vectors, with the possibility of using threshold requirements and that of averaging the different vectors. Although the Bayesian framework and some probabilistic assumptions provide useful indications about the potential advantage of different combination schemes, the combined performance ultimately depends on the joint probability distribution of the outputs, and it can be estimated by joining the results of different nets on the same test set. The combination methods are very flexible, they permit a straightforward cooperation of neural and traditional recognizers, and they are appropriate in a development environment where experiments are performed with different kinds of nets and features for a selected application. From the authors' experiments in the field of handwritten digit recognition (up to a total of more than 50000 characters), they found that the use of a small number of nets (two to three) with a sufficiently large uncorrelation in their mistakes reaches a combined performance that is significantly higher than the best obtainable from the individual nets, with a negligible effort after starting from a pool of networks produced in the development phase of an application. In particular, for a real-world OCR application, the best accuracy increase is about half the increase in the rejection rate, so that accuracies of the order of 99.5% can be reached by rejecting less than 5% of the patterns. This performance is significant for real applications.}, } @ARTICLE{bauer99empirical, author = {E. Bauer and R. Kohavi}, title = {An Empirical Comparison of Voting Classification Algorithms: Bagging, Boosting, and Variants}, journal = {Machine Learning}, year = 1999, volume = 36, number = {1,2}, } @ARTICLE{bishop95training, author = {Chris M. Bishop}, title = {Training with Noise is Equivalent to {T}ikhonov Regularization}, journal = {Neural Computation}, volume = 7, number = 1, pages = {108--116}, year = 1995, url = {http://citeseer.nj.nec.com/bishop94training.html}, } @BOOK{bishop95book, author = {Christopher M. Bishop}, title = {Neural Networks for Pattern Recogntion}, publisher = {Oxford University Press}, year = 1995, isbn = {0-19-853864-2}, } @MISC{uci, author = {C.L. Blake and C.J. Merz}, year = 1998, title = {{UCI} Repository of machine learning databases}, url = {http://www.ics.uci.edu/$\sim$mlearn/MLRepository.html}, institution = {University of California, Irvine, Dept. of Information and Computer Sciences}, } @INPROCEEDINGS{bottou91framework, author = {L{\'e}on Bottou and Patrick Gallinari}, title = {A Framework for the Cooperation of Learning Algorithms}, booktitle = {Advances in Neural Information Processing Systems}, volume = 3, publisher = {Morgan Kaufmann Publishers, Inc.}, editor = {Richard P. Lippmann and John E. Moody and David S. Touretzky}, pages = {781--788}, year = 1991, } @TECHREPORT{breiman96arcing, author = {Leo Breiman}, title = {Bias, Variance, and Arcing Classifiers}, institution = {Statistics Department, Berkeley}, year = 1996, number = 460, } @ARTICLE{breiman96bagging, author = {Leo Breiman}, title = {Bagging Predictors}, journal = {Machine Learning}, volume = 24, number = 2, pages = {123-140}, year = 1996, } @TECHREPORT{breiman98randomizing, author = {Leo Breiman}, title = {Randomizing Outputs to increase prediction accuracy}, institution = {Statistics Department, University of California}, month = {May}, year = 1998, type = {Technical Report}, number = 518, url = {http://www.boosting.org/papers/Bre98.pdf}, } @TECHREPORT{breiman99random, author = {Leo Breiman}, year = 1999, title = {Random Forests Random Features}, institution = {University of California, Berkley (Dept of Statistics)}, number = 567, } @INPROCEEDINGS{brodleylane96, author = {C. Brodley and T. Lane}, title = {Creating and exploiting coverage and diversity}, booktitle = {AAAI-96 Workshop Integrating Multiple Learned Models}, year = 1996, } @TECHREPORT{carneytuning99, author = {John Carney and Padraig Cunningham}, title = {Tuning diversity in bagged neural network ensembles}, institution = {Trinity College Dublin}, year = 1999, number = {TCD-CS-1999-44}, } @INPROCEEDINGS{chan95arbiter, author = {Philip K.~Chan and Salvatore J.~Stolfo}, title = {Learning Arbiter and Combiner Trees from Partitioned Data for Scaling Machine Learning}, booktitle = {The first international conference on knowledge discovery and data mining, KDD '95}, year = 1995, pages = {39-45}, } @INPROCEEDINGS{chan96local, author = {Philip K.~Chan and Salvatore J.~Stolfo}, title = {Sharing Learned Models among Remote Database Partitions by Local Meta-learning}, booktitle = {Proc. Second Intl. Conf. on Knowledge Discovery \& Data Mining}, year = 1996, pages = {2--7}, } @ARTICLE{chan98accuracy, author = {Philip K.~Chan and Salvatore J.~Stolfo}, title = {On the Accuracy of Meta-learning for Scalable Data Mining,}, journal = {Journal of Intelligent Information Systems}, year = 1997, volume = 8, pages = {5--28}, } @ARTICLE{clemen89forecasts, author = {R. Clemen}, year = 1989, title = {Combining forecast: A review and annotated bibliography}, journal = {International Journal on Forecasting}, volume = 5, pages = {559--583}, } @INPROCEEDINGS{costa95bayesian, author = {M. Costa and E. Filippi and E. Pasero}, title = {Artificial neural network ensembles: a Bayesian standpoint}, booktitle = {Proceedings of the 7th Italian Workshop on Neural Nets}, pages = {39-57}, publisher = {World Scientific}, year = 1995, editor = {M. Marinaro and R. Tagliaferri}, abstract = {Pooled estimates naturally arise from the Bayesian framework as the elected approach to both regression and classification problems whenever optimality in the average sense is concerned. In contrast, selection of a single 'best' member appears to be a somewhat crude approximation where some important features of the underlying model are discarded. However, the elegant and powerful full fledged formalism carries a very high computational complexity. Some practical implementations relevant to artificial neural network ensembles are therefore reviewed that make things more tractable while keeping the same generality.}, } @ARTICLE{hoeting99bma, author = {D. Hoeting, J. A.and Madigan and C.T. Raftery, A.E.and Volinsky}, title = {Bayesian model averaging: A tutorial}, journal = {Statistical Science}, year = 1999, volume = 44, number = 4, pages = {382--417}, url = {http://www.stat.washington.edu/www/research/online/hoeting1999.pdf}, url = {http://citeseer.nj.nec.com/context/1052498/0}, } @ARTICLE{darwenyao97, author = {Paul J. Darwen and Xin Yao}, title = {Speciation as Automatic Categorical Modularization}, journal = {{IEEE} {T}rans. on {E}volutionary {C}omputation}, volume = 1, number = 2, pages = {100--108}, year = 1997, } @TECHREPORT{darwenyao95, year = 1995, author = {Paul Darwen and Xin Yao}, institution = {University of New South Wales}, number = {CS 8/95}, title = {How Good is Fitness Sharing with a Scaling Function}, month = {April}, } @INCOLLECTION{darwenyao96, publisher = {Springer-Verlag}, author = {Paul Darwen and Xin Yao}, title = {Every niching method has its niche: fitness sharing and implicit sharing compared}, year = 1996 , booktitle = {Proc. of Parallel Problem Solving from Nature (PPSN) IV - Lecture Notes in Computer Science 1141}, } @ARTICLE{deb99multiobjective, author = {Kalyanmoy Deb}, title = {Multi-objective Genetic Algorithms: Problem Difficulties and Construction of Test Problems}, journal = {Evolutionary Computation}, volume = 7, number = 3, pages = {205-230}, year = 1999, } @ARTICLE{devroye79distribution, author = {L. Devroye and T. Wagner}, title = {Distribution-free Performance Bounds for Potential Function Rules}, journal = {IEEE Transactions on Information Theory}, year = 1979, volume = 25, number = 5, pages = {601--604}, } @INPROCEEDINGS{dietterichbakiri91:errorcorrecting, author = {T. G. Dietterich and G. Bakiri}, title = {Error-correcting output codes: a general method for improving multiclass inductive learning programs}, booktitle = {Proceedings of the Ninth {AAAI} National Conference on Artificial Intelligence}, publisher = {AAAI Press}, address = {Menlo Park, CA}, editor = {Dean, T. L. and McKeown, K.}, pages = {572--577}, year = 1991, } @ARTICLE{dietterich98approximate, author = {Thomas G. Dietterich}, title = {Approximate Statistical Test For Comparing Supervised Classification Learning Algorithms}, journal = {Neural Computation}, volume = 10, number = 7, pages = {1895-1923}, year = 1998, url = {citeseer.nj.nec.com/dietterich98approximate.html}, abstract = {This paper reviews five approximate statistical tests for determining whether one learning algorithm out-performs another on a particular learning task. These tests are compared experimentally to determine their probability of incorrectly detecting a difference when no difference exists (type I error). Two widely-used statistical tests are shown to have high probability of Type I error in certain situations and should never be used. These tests are (a) a test for the difference of two proportions and (b) a paired-differences t test based on taking several random train/test splits. A third test, a paired-differences t test based on 10-fold cross-validation, exhibits somewhat elevated probability of Type I error. A fourth test, McNemar's test, is shown to have low Type I error. The fifth test is a new test, 5x2cv, based on 5 iterations of 2-fold cross-validation. Experiments show that this test also has acceptable Type I error.}, } @INPROCEEDINGS{domingos97, title = {Why Does Bagging Work? {A} Bayesian Account and its Implications}, author = {P.~Domingos}, pages = 155, booktitle = {Proceedings of the Third International Conference on Knowledge Discovery and Data Mining ({KDD}-97)}, year = 1997, editor = {David Heckerman and Heikki Mannila and Daryl Pregibon and Ramasamy Uthurusamy}, publisher = {AAAI Press}, url = {http://www.boosting.org/papers/Dom97.ps.gz}, ps = {http://www.boosting.org/papers/Dom97.ps}, psgz = {http://www.boosting.org/papers/Dom97.ps.gz}, pdf = {http://www.boosting.org/papers/Dom97.pdf}, } @INPROCEEDINGS{domingos98occams, author = {Pedro Domingos}, title = {Occam's Two Razors: The Sharp and the Blunt}, booktitle = {Proceedings of the Fourth International Conference on Knowledge Discovery and Data Mining}, publisher = {AAAI Press}, year = 1998, } @ARTICLE{drucker94boosting, author = {Harris Drucker and Corinna Cortes and L. D. Jackel and Yann LeCun and Vladimir Vapnik}, title = {Boosting and Other Ensemble Methods}, type = {Letter}, journal = {Neural Computation}, volume = 6, number = 6, pages = {1289--1301}, year = 1994, abstract = {Compares the performance of three types of neural network-based ensemble techniques to that of a single neural network. The ensemble algorithms are two versions of boosting and committees of neural networks trained independently. For each of the four algorithms, we experimentally determine the test and training error curves in an optical character recognition (OCR) problem as both a function of training set size and computational cost, using three architectures. We show that a single machine is best for small training set sizes, while for large training set sizes, some version of boosting is best. However, for a given computational cost, boosting is always best. Furthermore, we show a surprising result for the original boosting algorithm, namely that as the training set size increases, the training error decreases until it asymptotes to the test error rate. This has potential implications in the search for better training algorithms. }, } @INCOLLECTION{edelmanintrator97, author = {S. Edelman and N. Intrator}, title = {Learning as Extraction of Low-Dimensional Representations}, booktitle = {Mechanisms of Perceptual Learning}, publisher = {Academic Press}, year = 1997, editor = {D. Medlin, R. Goldstone, P. Schyns}, } @BOOK{efron93bootstrap, author = {B. Efron and R. Tibshirani}, title = {An Introduction to the Bootstrap}, publisher = {Chapman and Hall}, year = 1993, } @ARTICLE{elkan1997ban, title = {{Boosting and Naive Bayesian Learning}}, author = {Elkan, C.}, journal = {proceeding of KDD-97, New Port beach, CA}, year = 1997 } @INPROCEEDINGS{fan99adacost, author = {Wei Fan and Salvatore J. Stolfo and Junxin Zhang and Philip K. Chan}, title = {Ada{C}ost: misclassification cost-sensitive boosting}, booktitle = {Proc. 16th International Conf. on Machine Learning}, publisher = {Morgan Kaufmann, San Francisco, CA}, pages = {97--105}, year = 1999, url = {http://citeseer.nj.nec.com/fan99adacost.html}, } @ARTICLE{feder94entropy, author = {Meir Feder and Neri Merhav}, title = {Relations between entropy and error probability}, journal = {IEEE Transactions on Information Theory}, volume = 40, number = 1, year = 1994, pages = 259, } @INPROCEEDINGS{feraud98ensemblemodular, author = {Rapha{\"e}l Feraud and Olivier Bernier}, title = {Ensemble and Modular Approaches for Face Detection: {A} Comparison}, booktitle = {Advances in Neural Information Processing Systems}, volume = 10, year = 1998, publisher = {The {MIT} Press}, editor = {Michael I. Jordan and Michael J. Kearns and Sara A. Solla}, } @ARTICLE{freund99short, author = {Y. Freund and R. Schapire}, title = {A short introduction to boosting}, journal = {Journal of Japanese Society for Artificial Intelligence}, year = 1999, pages = {771--780}, volume = 14, number = 5, url = {http://citeseer.nj.nec.com/freund99short.html}, } @INPROCEEDINGS{freundschapire96experiments, author = {Yoav Freund and Robert E. Schapire}, title = {Experiments with a new boosting algorithm}, booktitle = {Proceedings of the 13th International Conference on Machine Learning}, publisher = {Morgan Kaufmann}, year = 1996, pages = {148--156}, abstract = {In an earlier paper, we introduced a new 'boosting' algorithm called AdaBoost which, theoretically, can be used to significantly reduce the error of any learning algorithm that consistently generates classifiers whose performance is a little better than random guessing. We also introduced the related notion of a 'pseudo-loss' which is a method for forcing a learning algorithm of multi-label concepts to concentrate on the labels that are hardest to discriminate. In this paper, we describe experiments we carried out to assess how well AdaBoost with and without pseudo-loss, performs on real learning problems. We performed two sets of experiments. The first set compared boosting to Breiman's 'bagging' method when used to aggregate various classifiers (including decision trees and single attribute-value tests). We compared the performance of the two methods on a collection of machine-learning benchmarks. In the second set of experiments, we studied in more detail the performance of boosting using a nearest-neighbor classifier on an OCR problem.}, } @TECHREPORT{friedman99bagging, author = {J. Friedman and P. Hall}, title = {On Bagging and Nonlinear Estimation - available online at http://citeseer.nj.nec.com/friedman99bagging.html}, institution = {Stanford University}, year = 1999, url = {http://citeseer.nj.nec.com/friedman99bagging.html}, } @ARTICLE{friedman91mars, author = {J.H. Friedman}, title = {Multivariate Adaptive Regression Splines}, journal = {Annals of Statistics}, year = 1991, volume = 19, pages = {1--141}, } @TECHREPORT{friedman96, author = {J.H. Friedman}, title = {Bias, Variance, 0-1 Loss and the Curse of Dimensionality}, institution = {Stanford University}, year = 1996, } @ARTICLE{geman92, author = {S. Geman and E. Bienenstock and R. Doursat}, title = {Neural Networks and the Bias/Variance Dilemma}, type = {Letter}, journal = {Neural Computation}, volume = 4, number = 1, pages = {1--58}, year = 1992, abstract = {Feedforward neural networks trained by error backpropagation are examples of nonparametric regression estimators. We present a tutorial on nonparametric inference and its relation to neural networks, and we use the statistical viewpoint to highlight strengths and weaknesses of neural models. We illustrate the main points with some recognition. experiments involving artificial data as well as handwritten numerals. In way of conclusion, we suggest that current-generation feedforward neural networks are largely inadequate for difficult problems in machine perception and machine learning, regardless of parallel-versus-serial hardware or other implementation issues. Furthermore, we suggest that the fundamental challenges in neural modeling are about representation rather than learning per se. This last point is supported by additional experiments with handwritten numerals.}, } @ARTICLE{girosi95regularization, author = {Federico Girosi and Michael Jones and Tomaso Poggio}, title = {Regularization Theory and Neural Networks Architectures}, journal = {Neural Computation}, volume = 7, number = 2, pages = {219--269}, year = 1995, url = {http://citeseer.nj.nec.com/girosi95regularization.html}, } @ARTICLE{granger89combining, pages = {167--174}, author = {C.W.J. Granger}, year = 1989, journal = {Journal of Forecasting}, volume = 8, title = {Combining Forecasts -- Twenty Years Later}, } @INPROCEEDINGS{guerrasalcedo99genetic, author = {Cesar Guerra-Salcedo and Darrell Whitley}, title = {Genetic Approach to Feature Selection for Ensemble Creation}, booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference}, volume = 1, month = {13-17}, publisher = {Morgan Kaufmann}, address = {Orlando, Florida, USA}, editor = {Wolfgang Banzhaf and Jason Daida and Agoston E. Eiben and Max H. Garzon and Vasant Honavar and Mark Jakiela and Robert E. Smith}, isbn = {1-55860-611-4}, pages = {236--243}, year = 1999, url = {citeseer.ist.psu.edu/531553.html}, } @ARTICLE{hansen99combining, author = {J.V. Hansen}, title = {Combining Predictors: Comparison of Five Meta Machine Learning Methods}, journal = {Information Sciences}, volume = 119, number = {1-2}, pages = {91-105}, year = 1999, } @ARTICLE{hansensalamon90, volume = 12, number = 10, title = {Neural Network Ensembles}, author = {Lars Kai Hansen and Peter Salamon}, journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence}, year = 1990, pages = {993-1001}, } @MISC{haselsteiner99dynamic, author = {E. Haselsteiner}, title = {Dynamic targets - adapting supervised learning to time series classification}, text = {In Proceedings of the International Joint Conference on Neural Networks IJCNN'99, Washington D.C., IEEE Press.}, year = 1999, } @PHDTHESIS{hashem93thesis, author = {Sherif Hashem}, title = {{O}ptimal {L}inear {C}ombinations of {N}eural {N}etworks}, school = {School of Industrial Engineering, University of Purdue}, year = 1993, } @ARTICLE{hashem97optimal, author = {Sherif Hashem}, title = {{O}ptimal {L}inear {C}ombinations of {N}eural {N}etworks}, journal = {Neural Networks}, volume = 10, number = 4, month = {August}, pages = {599-614}, year = 1997, } @ARTICLE{hashem95optimal, author = {Sherif Hashem and Bruce Schmeiser}, title = {Improving Model Accuracy Using Optimal Linear Combinations of Trained Neural Networks}, journal = {IEEE Transactions on Neural Networks}, type = {Letter}, year = 1995, volume = 6, number = 3, pages = {792--794}, month = may, abstract = {Neural network (NN) based modeling often requires trying multiple networks with different architectures and training parameters in order to achieve an acceptable model accuracy. Typically, only one of the trained networks is selected as 'best' and the rest are discarded. The authors propose using optimal linear combinations (OLC's) of the corresponding outputs on a set of NN's as an alternative to using a single network. Modeling accuracy is measured by mean squared error (MSE) with respect to the distribution of random inputs. Optimality is defined by minimizing the MSE, with the resultant combination referred to as MSE-OLC. The authors formulate the MSE-OLC problem for trained NN's and derive two closed-form expressions for the optimal combination-weights. An example that illustrates significant improvement in model accuracy as a result of using MSE-OLC's of the trained networks is included.}, } @ARTICLE{ho98subspaces, author = {T.K. Ho}, title = {The Random Subspace Method for Constructing Decision Forests}, journal = {IEEE Trans. on Pattern Analysis and Machine Intelligence}, year = 1998, volume = 20, number = 8, pages = {832--844}, month = {August}, } @ARTICLE{ho94mcs, author = {Tin Kam Ho and Jonathan J.~Hull and Sargur N.~Srihari}, title = {Decision Combination in Multiple Classifier Systems}, journal = {Pattern Analysis and Machine Intelligence}, year = 1994, volume = 16, number = 1, pages = {66--75}, month = {January}, } @INPROCEEDINGS{huang94nn, author = {Y S Huang and C Y Suen}, title = {A Method of Combining Multiple Classifiers - A Neural Network Approach}, booktitle = {Proceedings of the 12th International Conference on Pattern Recognition and Computer Vision}, year = 1994, pages = {473-475}, address = {Jerusalem, Israel}, } @ARTICLE{huang95numerals, author = {Y.S.~Huang and C.Y.~Suen}, title = {A Method of Combining Multiple Experts for the Recognition of Unconstrained Handwritten Numerals}, journal = {Pattern Analysis and Machine Intelligence}, year = 1995, volume = 17, number = 1, pages = {90--94}, month = {January}, } @ARTICLE{husmeier98overfitting, author = {Husmeier D., Althoefer K.}, title = {Modelling conditional probabilities with network committees: how overfitting can be useful}, journal = {Neural Network World}, year = 1998, volume = 8, numver = 4, pages = {417--439}, } @ARTICLE{intratoredelman96, author = {N. Intrator and S. Edelman}, title = {Making a Low-Dimensional Representation Suitable for Diverse Tasks}, journal = {Connection Science : Special Issue on Reuse of Neural Networks Through Transfer}, year = 1996, volume = 8, number = 2, pages = {205-224}, } @ARTICLE{jacobsjordan91, author = {R. A. Jacobs and M. I. Jordan and S. J. Nowlan and G. E. Hinton}, title = {Adaptive Mixtures of Local Experts}, type = {Letter}, journal = {Neural Computation}, volume = 3, number = 1, pages = {79--87}, year = 1991, abstract = {We present a new supervised learning procedure for systems composed of many separate networks, each of which learns to handle a subset of the complete set of training cases. The new procedure can be viewed either as a modular version of a multilayer supervised network, or as an associative version of competitive learning. It therefore provides a new link beetween these two apparently different approaches. We demonstrate that the learning procedure divides up a voewel discrimination task into appropriate subtasks, each of which can be solved by a very simple expert network.}, } @INBOOK{jacobs99mixturesofx, author = {R.A. Jacobs and M.A. Tanner}, chapter = {Mixtures of {X}}, editor = {A.J. Sharkey}, title = {Combining Articial Neural Nets}, publisher = {Springer-Verlag, London}, year = 1999, } @ARTICLE{jacobs97analyses, pages = {369--383}, author = {Robert Jacobs}, year = 1997, journal = {Neural Computation}, volume = 9, title = {Bias-Variance Analyses of Mixture-of-Experts Architectures}, } @ARTICLE{jacobsjordanbarto91, pages = {219-250}, title = {Task decomposition through competition in a modular connectionist architecture - the What and Where vision tasks}, volume = 15, year = 1991, author = {Robert A. Jacobs and Michael I. Jordan and Andrew G. Barto}, journal = {Cognitive Science}, } @INPROCEEDINGS{jelonek96replicated, author = {J. Jelonek}, title = {Generalization Capability of Homogeneous Voting Classifier Based on Partially Replicated Data}, series = {Integrating Multiple Learned Models for Improving and Scaling Machine Learning Algorithms Workshop}, booktitle = {AAAI'96}, year = 1996, note = {Portland, OR}, } @ARTICLE{jima97weak, author = {C. Ji and S. Ma}, title = {Combinations of Weak Classifiers}, journal = {IEEE Trans. on Neural Networks, Special Issue on Neural Networks and Pattern Recognition}, volume = 8, year = 1997, month = {January}, pages = {32--42}, } @INPROCEEDINGS{jimenez98dynamically, author = {D. Jimenez and N. Walsh}, title = {Dynamically weighted ensemble neural networks for classification}, booktitle = {Proceedings of the International Joint Conference on Neural Networks}, year = 1998, url = {http://citeseer.nj.nec.com/jimenez98dynamically.html}, } @ARTICLE{jordan99introduction, author = {Michael I. Jordan and Zoubin Ghahramani and Tommi Jaakkola and Lawrence K. Saul}, title = {An Introduction to Variational Methods for Graphical Models}, journal = {Machine Learning}, volume = 37, number = 2, pages = {183-233}, year = 1999, } @ARTICLE{jordan94hierarchical, author = {Michael I. Jordan and Robert A. Jacobs}, title = {Hierarchical Mixtures of Experts and the {EM} Algorithm}, journal = {Neural Computation}, year = 1994, volume = 6, pages = {181--214}, class = {nn, learning}, abstract = {We present a tree-structured architecture for supervised learning. The statistical model underlying the architecture is a hierarchical mixture model in which both the mixture coefficients and the mixture components are generalized linear models (GLIM). Learning is treated as a maximum likelihood problem; in particular, we present an Expectation-Maximization (EM) algorithm for adjusting the parameters of the architecture. We develop an online learning algorithm in which the parameters are updated incrementally. Comparative simulation results are presented in the robot dynamics domain.}, } @INPROCEEDINGS{kang95learning, author = {Kukjin Kang and Jong-Hoon Oh}, title = {Learning by a Population of Perceptrons}, booktitle = {Computational Learning Theory}, pages = {297-300}, year = 1995, url = {http://citeseer.nj.nec.com/kang97learning.html}, } @INPROCEEDINGS{kang97statistical, author = {Kukjin Kang and Jong-Hoon Oh}, title = {Statistical Mechanics of the Mixture of Experts}, booktitle = {Advances in Neural Information Processing Systems}, volume = 9, publisher = {The {MIT} Press}, editor = {Michael C. Mozer and Michael I. Jordan and Thomas Petsche}, pages = 183, year = 1997, url = {http://citeseer.nj.nec.com/kang97statistical.html}, } @ARTICLE{kleinberg90stochastic, author = {E. M. Kleinberg}, title = {Stochastic Discrimination}, journal = {Annals of Mathematics and Artificial Intelligence}, volume = 1, year = 1990, } @INPROCEEDINGS{kleinberg93pattern, author = {E. M. Kleinberg and T. K. Ho}, title = {Pattern Recognition by Stochastic Modeling}, booktitle = {Proc. of the 3rd Workshop on Frontiers in Handwriting Recognition}, month = {May}, address = {Buffalo, New York}, pages = {175--183}, year = 1993, } @INPROCEEDINGS{kohavi96bias, author = {Ron Kohavi and David H. Wolpert}, title = {Bias Plus Variance Decomposition for Zero-One Loss Functions}, booktitle = {Machine Learning: Proceedings of the Thirteenth International Conference}, publisher = {Morgan Kaufmann}, editor = {Lorenza Saitta}, pages = {275--283}, year = 1996, abstract = {We present a bias-variance decomposition of expected misclassification rate, the most commonly used loss function in supervised classification learning. The bias-variance decomposition for quadratic loss functions is well known and serves as an important tool for analyzing learning algorithms, yet no decomposition was offered for the more commonly used zero-one (misclassification) loss functions until the work of Kong and Dietterich (1995) and Breiman (1996). Their decomposition suffers from some major shortcomings though (e.g., potentially negative variance), which our decomposition avoids. We show that, in practice, the naive frequency-based estimation of the decomposition terms is by itself biased and show how to correct for this bias. We illustrate the decomposition on various algorithms and datasets from the UCI repository.}, } @INPROCEEDINGS{kongdietterich95correcting, author = {E. B. Kong and T. G. Dietterich}, title = {Error-Correcting Output Coding Corrects Bias and Variance}, booktitle = {Proceedings of the 12th International Conference on Machine Learning}, publisher = {Morgan Kaufmann}, year = 1995, pages = {313--321}, comment = {Tahoe Cite, CA}, abstract = {Previous research has shown that a technique called error-correcting output coding (ECOC) can dramatically improve the classification accuracy of supervised learning algorithms that learn to classify data points into one of k>>2 classes. This paper presents an investigation of why the ECOC technique works, particularly when employed with decision-tree learning algorithms. It shows that the ECOC method, like any form of voting or committee, can reduce the variance of the learning algorithm. Furthermore, unlike methods that simply combine multiple runs of the same learning algorithm, ECOC can correct errors caused by the bias of the learning algorithm. Experiments show that this bias correction ability relies on the non-local behavior of C4.5.}, } @ARTICLE{kovalishyn98cascade, author = {Kovalishyn, V. V. and Tetko, I. V. and Luik, A. I. and Kholodovych, V. V. and Villa, A. E. P. and Livingstone, D. J.}, title = {Neural network studies. 3. Variable selection in the cascade-correlation learning architecture}, journal = {Journal of Chemical Information & Computer Sciences}, volume = 38, number = 4, pages = {651-659}, abstract = {Pruning methods for feed-forward artificial neural networks trained by the cascade-correlation learning algorithm are proposed. The cascade-correlation algorithm starts with a small network and dynamically adds new nodes until the analyzed problem has been solved. This feature of the algorithm removes the requirement to predefine the architecture of the neural network prior to network training. The developed pruning methods are used to estimate the importance of large sets of initial variables for quantitative structure-activity relationship studies and simulated data sets. The calculated results are compared with the performance of fixed-size back-propagation neural networks and multiple regression analysis and are carefully validated using different training/test set protocols, such as leave-one-out and full cross-validation procedures. The results suggest that the pruning methods can be successfully used to optimize the set of variables for the cascade-correlation learning algorithm neural networks. The use of variables selected by the elaborated methods provides an improvement of neural network prediction ability compared to that calculated using the unpruned sets of variables.}, year = 1998, } @BOOK{koza92book, author = {John R. Koza}, title = {Genetic Programming: On the Programming of Computers by Means of Natural Selection}, publisher = {MIT Press}, year = 1992, isbn = {0-262-11170-5}, } @ARTICLE{kroghsollich97:statistical_mechanics, title = {Statistical mechanics of ensemble learning}, author = {A. Krogh and P. Sollich}, journal = {Physical Review E}, year = 1997, volume = 55, number = {1PtB}, pages = {811--825}, abstract = {Within the context of learning a rule from examples, we study the general characteristics of learning, with ensembles. The generalization performance achieved by a simple model ensemble of linear students is calculated exactly in the thermodynamic limit of a large number of input components and shows a surprisingly rich behavior. Our main findings are the following. For learning in large ensembles, it is advantageous to use underregularized students, which actually overfit the training data. Globally optimal generalization performance can be obtained by choosing the training set sizes of the students optimally. For smaller ensembles, optimization of the ensemble weights can yield significant improvements in ensemble generalization performance, in particular if the individual students are subject to noise in the training process. Choosing students with a wide range of regularization parameters makes this improvement robust against changes in the unknown level of corruption of the training data.}, } @ARTICLE{kroghvedelsby95, author = {A. Krogh and J. Vedelsby}, title = {Neural Network Ensembles, Cross Validation, and Active Learning}, journal = {NIPS}, year = 1995, volume = 7, pages = {231--238}, abstract = {The learning of continuous-valued functions using neural network ensembles (committees) can give improved accuracy, a reliable estimation of the generalization error, and active learning. The ambiguity is defined as the variation of the output of ensemble members averaged over unlabeled data, so it quantifies the disagreement among the networks. We discuss how to use the ambiguity in combination with cross-validation to give a reliable estimate of the ensemble generalization error, and how this type of ensemble cross-validation can sometimes improve performance. We show how to estimate the optimal weights of the ensemble members using unlabeled data. By a generalization of query-by-committee, we show how the ambiguity can be used to select new training data to be labeled in an active learning scheme.}, } @BOOK{laplace1818, author = {P. S. de Laplace}, title = {Deuxieme supplement a la theorie analytique des probabilites}, publisher = {Paris, Gauthier-Villars}, year = 1818, note = {Reprinted (1847) in Oeuvres Completes de Laplace, vol. 7}, annote = {First known observation that combining probabilistic classifiers can help.}, } @ARTICLE{leblanc96combining, author = {Michael LeBlanc and Robert Tibshirani}, title = {Combining Estimates in Regression and Classification}, journal = {Journal of the American Statistical Association}, volume = 91, number = 436, pages = 1641, year = 1996, url = {http://citeseer.nj.nec.com/leblanc93combining.html}, } @ARTICLE{lehto95, volume = 7, author = {Mikko Lehtokangas and Jukka Saarinen and Kimmo Kaski and Pentti Huuhtanen}, year = 1995, pages = {983-999}, number = 5, title = {Initializing Weights of a Multilayer Perceptron Network by Using the Orthogonal Least Squares Algorithm}, journal = {Neural Computation}, } @ARTICLE{liaomoody99, author = {Yuansong Liao and John Moody}, title = {Constructing Heterogeneous Committees Using Input Feature Grouping}, journal = {Advances in Neural Information Processing Systems}, year = 1999, volume = 12, } @PHDTHESIS{liu98thesis, author = {Y. Liu}, title = {Negative Correlation Learning and Evolutionary Neural Network Ensembles}, school = {University College, The University of New South Wales, Australian Defence Force Academy, Canberra, Australia}, year = 1998, } @ARTICLE{liuyao97negatively, title = {Negatively Correlated Neural Networks can Produce Best Ensembles}, author = {Y. Liu and X. Yao}, number = {3/4}, year = 1997, journal = {Australian Journal of Intelligent Information Processing Systems}, pages = {176--185}, volume = 4, } @INPROCEEDINGS{liuyao98towards, month = {February}, author = {Yong Liu and Xin Yao}, year = 1998 , booktitle = {Proceedings of International Symposium on Artificial Life and Robotics (AROB)}, pages = {265-268}, title = {Towards Designing Neural Network Ensembles by Evolution}, } @ARTICLE{liuyao99:ensemblelearningvia, author = {Yong Liu and Xin Yao}, title = {Ensemble learning via negative correlation}, journal = {Neural Networks}, volume = 12, number = 10, pages = {1399--1404}, year = 1999, } @INPROCEEDINGS{maclinshavlik95weights, author = {R. Maclin and J. W. Shavlik}, title = {Combining the Predictions of Multiple Classifiers: Using Competitive Learning to Initialize Neural Networks}, booktitle = {Proceedings of the 14th International Joint Conference on Artificial Intelligence, Montreal, Canada}, year = 1995, pages = {524-530}, abstract = {The primary goal of inductive learning is to generalize well-that is, induce a function that accurately produces the correct output for future inputs. Hansen and Salamon (1990) showed that, under certain assumptions, combining the predictions of several separately trained neural networks will improve generalization. One of their key assumptions is that the individual networks should be independent in the errors they produce. In the standard way of performing backpropagation this assumption may be violated, because the standard procedure is to initialize network weights in the region of weight space near the origin. This means that backpropagation's gradient-descent search may only reach a small subset of the possible local minima. In this paper we present an approach to initializing neural networks that uses competitive learning to intelligently create networks that are originally located far from the origin of weight space, thereby potentially increasing the set of reachable local minima. We report experiments on two real-world datasets where combinations of networks initialized with our method generalize better than combinations of networks initialized the traditional way.}, } @INPROCEEDINGS{maclin97empirical, author = {Richard Maclin and David Opitz}, title = {An Empirical Evaluation of Bagging and Boosting}, booktitle = {{AAAI}/{IAAI}}, pages = {546-551}, year = 1997, url = {http://citeseer.nj.nec.com/maclin97empirical.html}, } @MISC{mak97combining, author = {B. Mak}, title = {Combining ANNs to improve phoneme recognition}, text = {B. Mak. Combining ANNs to improve phoneme recognition. ICASSP, 4:3253--3256, 1997.}, year = 1997, url = {http://citeseer.nj.nec.com/mak97combining.html}, } @INBOOK{mandler88:combining, author = {E. Mandler and J. Schuermann}, chapter = {Combining the Classification Results of independent classifiers based on the Dempster/Shafer theory of evidence}, title = {Pattern Recognition and Artificial Intelligence}, editor = {E.S.Gelsema and L.N.Kanal}, publisher = {North Holland, Amsterdam}, pages = {381--393}, year = 1988, } @ARTICLE{mani91portfolio, author = {G. Mani}, title = {Lowering variance of decisions by using artificial network portfolios}, type = {Letter}, journal = {Neural Computation}, volume = 3, number = 4, pages = {484--486}, year = 1991, } @INPROCEEDINGS{margineantu97pruning, author = {Dragos D. Margineantu and Thomas G. Dietterich}, title = {Pruning adaptive boosting}, booktitle = {Proc. 14th International Conference on Machine Learning}, publisher = {Morgan Kaufmann}, pages = {211--218}, year = 1997, url = {http://citeseer.nj.nec.com/margineantu97pruning.html}, } @ARTICLE{markowitz52, title = {Portfolio Selection}, author = {H. Markowitz}, journal = {Journal of Finance}, volume = 7, issue = 1, month = {March}, year = 1952, annote = {Shows that a linear combination (ensemble) of individual regressors obey the bias-variance-covariance decomposition. For this paper, Markowitz received the 1990 Nobel Prize for Economics.}, } @BOOK{mcclellandrumelhart86, author = {J. McClelland and D. Rumelhart}, title = {Parallel Distributed Processing}, publisher = {MIT Press}, year = 1986, } @ARTICLE{meir95bias, author = {Ronny Meir}, title = {Bias, Variance and the Combination of Least Squares Estimators}, pages = {295--302}, editor = {G. Tesauro and D. Touretzky and T. Leen}, volume = 7, booktitle = {Advances in Neural Information Processing Systems}, year = 1995, publisher = {The {MIT} Press}, abstract = {We consider the effect of combining several least squares estimators on the expected performance of a regression problem. Computing the exact bias and variance curves as a function of the sample size we are able to quantitatively compare the effect of the combination on the bias and variance separately, and thus on the expected error which is the sum of the two. Our exact calculations, demonstrate that the combination of estimators is particularly useful in the case where the data set is small and noisy and the function to be learned is unrealizable. For large data sets the single estimator produces superior results. Finally, we show that by splitting the data set into several independent parts and training each estimator on a different subset, the performance can in some cases be significantly improved.}, } @PHDTHESIS{merz98thesis, author = {C.J. Merz}, title = {Classification and Regression by Combining Models}, school = {University of California, Irvine}, year = 1998, abstract = {Two novel methods for combining predictors are introduced in this thesis; one for the task of regression, and the other for the task of classification. The goal of combining the predictions of a set of models is to form an improved predictor. This dissertation demonstrates how a combining scheme can rely on the stability of the consensus opinion and, at the same time, capitalize on the unique contributions of each model. An empirical evaluation reveals that the new methods consistently perform as well or better than existing combining schemes for a variety of prediction problems. The success of these algorithms is explained empirically and analytically by demonstrating how they adhere to a set of theoretical and heuristic guidelines. A byproduct of the empirical investigation is the evidence that existing combining methods fail to satisfy one or more of the guidelines defined. The new combining approaches satisfy these criteria by relying upon Singular Value Decomposition as a tool for filtering out the redundancy and noise in the predictions of the learn models, and for characterizing the areas of the example space where each model is superior. The SVD-based representation used in the new combining methods aids in avoiding sensitivity to correlated predictions without discarding any learned models. Therefore, the unique contributions of each model can still be discovered and exploited. An added advantage of the combining algorithms derived in this dissertation is that they are not limited to models generated by a single algorithm; they may be applied to model sets generated by a diverse collection of machine learning and statistical modeling methods.}, } @ARTICLE{merz99using, author = {Christopher J. Merz}, title = {Using Correspondence Analysis to Combine Classifiers}, journal = {Machine Learning}, volume = 36, number = {1-2}, pages = {33-58}, year = 1999, } @INPROCEEDINGS{merz97combining, author = {Christopher J. Merz and Michael J. Pazzani}, title = {Combining Neural Network Regression Estimates with Regularized Linear Weights}, booktitle = {Advances in Neural Information Processing Systems}, volume = 9, publisher = {The {MIT} Press}, editor = {Michael C. Mozer and Michael I. Jordan and Thomas Petsche}, pages = 564, year = 1997, } @INBOOK{minsky90multiple, author = {Marvin Minsky}, editor = {Patrick H. Winston}, title = {Artificial Intelligence at MIT : Expanding Frontiers}, chapter = {Logical vs. Analogical or Symbolic vs. Connectionist or Neat vs Scruffy}, publisher = {MIT Press}, year = 1990, volume = 1, annote = {Minsky comments on the benefits of combining multiple representations.} } @TECHREPORT{moerland99dynaboost, author = {P.~Moerland and E.~Mayoraz}, title = {DynaBoost: Combining Boosted Hypotheses in a Dynamic Way}, institution = {IDIAP}, year = 1999, number = {RR 99-09}, address = {Switzerland}, month = {May}, url = {http://www.boosting.org/papers/MoeMay99.pdf}, } @ARTICLE{montgomeryfriedman93, author = {D.C. Montgomery and D.J. Friedman}, title = {Prediction Using Regression Models with Multicollinear Predictor Variables}, journal = {IIE Transactions}, year = 1993, volume = 25, number = 3, pages = {73--85}, abstract = {Linear regression models are widely used for forecasting and prediction of new observations from the underlying modeled process. The article explores the use of regression models in this context when the regressor or predictor variables exhibit multicollinearity, or near-linear dependence. It is shown that multicollinearity can severely impact the predictive performance of a regression model and that biased estimation methods can be an effective countermeasure when multicollinearity is present. Several biased estimation methods are described and evaluated, including a new method for selecting the biasing parameter in ordinary ridge regression. A simulation study is performed to provide some guidelines for the choice of an estimation method. }, } @INPROCEEDINGS{murata98bias, author = {N. Murata}, title = {Bias of estimators and regularization terms}, booktitle = {Proceedings of Workshop on Information-Based Induction Sciences}, pages = {87--94}, year = 1998, month = {July}, url = {http://citeseer.nj.nec.com/murata98bias.html}, } @INBOOK{murphy97exploring, author = {Patrick M. Murphy and Michael J. Pazzani}, title = {Exploring the decision forest: an empirical investigation of Occam's razor in decision tree induction}, booktitle = {Computational Learning Theory and Natural Learning Systems}, volume = {IV: Making Learning Systems Practical}, publisher = {MIT Press}, pages = {171--187}, year = 1997, url = {http://citeseer.nj.nec.com/murphy94exploring.html}, } @BOOK{nilsson65, author = {N.J.Nilsson}, title = {Learning Machines: Foundations of Trainable Pattern-Classifying Systems}, publisher = {McGraw-Hill}, year = 1965, } @ARTICLE{naftaly97optimal, author = {Ury Naftaly and Nathan Intrator and David Horn}, year = 1997, volume = 8, title = {Optimal Ensemble Averaging of Neural Networks}, month = {May}, number = 3 , pages = {283--296}, journal = {Network}, } @INBOOK{ohkang97, author = {Jong-Hoon Oh and Kookjin Kang}, editor = {K.Y.M.Wong and D.Yan}, title = {Theoretical Aspects of Neural Computation: A Multidisciplinary Perspective}, chapter = {Experts or Ensemble? A statistical Mechanics of Multiple Neural Network Approaches}, publisher = {Springer, Heidelberg}, pages = {81--92}, year = 1997, } @MISC{opitz99genetic, author = {D. Opitz and J. Shavlik}, title = {A genetic algorithm approach for creating neural network ensembles}, text = {In A.J.C. Sharkey, editor, Combining Articial Neural Nets, pages 79-99. Springer-Verlag, London, 1999.}, year = 1999, } @INPROCEEDINGS{opitz99feature, author = {David Opitz}, title = {Feature Selection for Ensembles}, booktitle = {Proceedings of 16th National Conference on Artificial Intelligence (AAAI)}, pages = {379-384}, year = 1999, } @ARTICLE{opitzmaclin99:popular, title = {Popular Ensemble Methods: An Empirical Study}, journal = {Journal of Artificial Intelligence Research}, volume = 11 , year = 1999, pages = {169-198}, author = {David Opitz and Richard Maclin}, } @ARTICLE{opitz96:generating, author = {David W. Opitz and Jude W. Shavlik}, title = {Generating Accurate and Diverse Members of a Neural-Network Ensemble}, pages = {535--541}, journal = {NIPS}, editor = {David S. Touretzky and Michael C. Mozer and Michael E. Hasselmo}, volume = 8, year = 1996, publisher = {The {MIT} Press}, abstract = {Neural-network ensembles have been shown to be very accurate classification techniques. Previous work has shown that an effective ensemble should consist of networks that are not only highly correct, but ones that make their errors on different parts of the input space as well. Most existing techniques, however, only indirectly address the problem of creating such a set of networks. In this paper we present a technique called ADDEMUP that uses genetic algorithms to directly search for an accurate and diverse set of trained networks. ADDEMUP works by first creating an initial population, then uses genetic operators to continually create new networks, keeping the set of networks that are as accurate as possible while disagreeing with each other as much as possible. Experiments on three DNA problems show that ADDEMUP is able to generate a set of trained networks that is more accurate than several existing approaches. Experiments also show that ADDEMUP is able to effectively incorporate prior knowledge, if available, to improve the quality of its ensemble.}, } @INPROCEEDINGS{ormoneit96improved, author = {Dirk Ormoneit and Volker Tresp}, title = {Improved Gaussian Mixture Density Estimates Using Bayesian Penalty Terms and Network Averaging}, booktitle = {Advances in Neural Information Processing Systems}, volume = 8, publisher = {The {MIT} Press}, editor = {David S. Touretzky and Michael C. Mozer and Michael E. Hasselmo}, pages = {542--548}, year = 1996, } @MISC{orr95regularisation, author = {M. Orr}, title = {Regularisation in the Selection of Radial Basis Function Centres}, text = {M. J. L. Orr, Regularisation in the Selection of Radial Basis Function Centres, Neural Computation, Vol. 7, pp. 606-623, 1995.}, year = 1995, url = {citeseer.nj.nec.com/orr95regularisation.html}, } @TECHREPORT{tumer99decimated, author = {N. Oza and K. Tumer}, title = {Dimensionality Reduction through Classifier Ensembles}, institution = {NASA Ames Labs}, year = 1999, number = {NASA-ARC-IC-1999-126}, } @ARTICLE{partridge96network, author = {D. Partridge}, title = {Network Generalization Differences Quantified}, journal = {Neural Networks}, volume = 9, number = 2, pages = {263--271}, year = 1996, url = {http://citeseer.nj.nec.com/partridge94network.html}, } @ARTICLE{partridge96:engineering, author = {D. Partridge and W. B. Yates}, title = {Engineering Multiversion Neural-Net Systems}, journal = {Neural Computation}, volume = 8, number = 4, pages = {869--893}, year = 1996, } @INPROCEEDINGS{nips90:Pearlmutter-Rosenfeld, author = {Barak A. Pearlmutter and Ronald Rosenfeld}, title = {Chaitin-{K}olmogorov Complexity and Generalization in Neural Networks}, pages = {925--931}, url = {http://nips.djvuzone.org/djvu/nips03/0925.djvu}, booktitle = {Advances in Neural Information Processing Systems 3}, publisher = {Morgan Kaufmann}, year = 1991, } @INCOLLECTION{perronecooper93, author = {M. P. Perrone and L. N. Cooper}, title = {When networks disagree: Ensemble methods for hybrid neural networks}, editor = {R. J. Mammone}, booktitle = {Artificial Neural Networks for Speech and Vision}, address = {London}, pages = {126--142}, year = 1993, } @PHDTHESIS{perrone93improving, author = {M.P. Perrone}, title = {Improving Regression Estimation: Averaging Methods for Variance Reduction with Extensions to General Convex Measure Optimization}, year = 1993, school = {Brown University, Institute for Brain and Neural Systems}, } @ARTICLE{prechelt98automatic, author = {Lutz Prechelt}, title = {Automatic early stopping using cross validation: quantifying the criteria}, journal = {Neural Networks}, volume = 11, number = 4, pages = {761--767}, year = 1998, url = {http://citeseer.nj.nec.com/prechelt98automatic.html}, } @INPROCEEDINGS{powalka95, author = {R Powalka, N Sherkat, R Whitrow}, title = {Multiple recognizer combination topologies}, booktitle = {Proceedings of the Seventh Biannual Conference of the International Graphonomics Society}, pages = {128--129}, year = 1995, month = {August}, organization = {University of Western Ontario}, note = {ISBN 0-921121-14-8}, } @ARTICLE{rahman98, author = {A.F.R. Rahman and M.C. Fairhurst}, title = {An evaluation of multi-expert configurations for the recognition of handwritten numerals}, journal = {Pattern Recognition}, year = 1998, number = 9, pages = {1255--1273}, } @ARTICLE{intratorraviv96:noise, volume = 8, pages = {355-372}, author = {Yuval Raviv and Nathan Intrator}, year = 1996, title = {Bootstrapping with Noise: An Effective Regularisation Technique}, journal = {Connection Science}, } @ARTICLE{rogova94combining, author = {Galina Rogova}, title = {Combining the Results of Neural Network Classifiers}, journal = {Neural Networks}, year = 1994, volume = 7, number = 5, pages = {777--781}, class = {nn, learning}, } @ARTICLE{rosen96:decorrelated, volume = 8, number = {3 and 4}, title = {Ensemble Learning using Decorrelated Neural Networks}, author = {B.E. Rosen}, journal = {Connection Science - Special Issue on Combining Artificial Neural Networks: Ensemble Approaches}, year = 1996, pages = {373--384}, abstract = {We describe a decorrelation network training method for improving the quality of regression learning in 'ensemble' neural networks (NNs) that are composed of linear combinations of individual NNs. In this method, individual networks are trained by backpropagation not only to reproduce a desired output, but also to have their errors linearly decorrelated with the other networks. Outputs from the individual networks are then linearly combined to produce the output of the ensemble network. We demonstrate the performances of decorrelated network training on learning the 'three-parity' logic function, a noisy sine function and a one-dimensional non-linear function, and compare the results with the ensemble networks composed of independently trained individual networks (without decorrelation training). Empirical results show that when individual networks are forced to be decorrelated with one another the resulting ensemble NNs have lower mean squared errors than the ensemble networks having independently trained individual networks. This method is particularly applicable when there is insufficient data to train each individual network on disjoint subsets of training patterns.}, } @ARTICLE{ruck90bayesperceptron, author = {D. W. Ruck and S. K. Rogers and M. Kabrisky and M. E. Oxley and B. W. Suter}, title = {The multilayer perceptron ass an approximation to a Bayes optimal discrimant function}, journal = {IEEE Transactions on Neural Networks}, type = {Letter}, year = 1990, volume = 1, number = 4, pages = {296--298}, } @INPROCEEDINGS{saborin93digit, author = {Michael Saborin and Amar Mitiche and Danny Thomas and George Nagy}, title = {Classifier Combination for Hand-Printed Digit Recognition}, pages = {163--166}, booktitle = {Proceedings of the Second International Conference on Document Analysis and Recognition}, year = 1993, organization = {IEEE}, address = {Japan}, } @ARTICLE{schapiresinger99confidence, author = {R.E.~Schapire and Y.~Singer}, title = {Improved boosting algorithms using confidence-rated predictions}, journal = {Machine Learning}, year = 1999, volume = 37, number = 3, month = dec, pages = {297-336}, abstract = { We describe several improvements to Freund and Schapire's \adaboost\ boosting algorithm, particularly in a setting in which hypotheses may assign confidences to each of their predictions. We give a simplified analysis of AdaBoost in this setting, and we show how this analysis can be used to find improved parameter settings as well as a refined criterion for training weak hypotheses. We give a specific method for assigning confidences to the predictions of decision trees, a method closely related to one used by Quinlan. This method also suggests a technique for growing decision trees which turns out to be identical to one proposed by Kearns and Mansour. We focus next on how to apply the new boosting algorithms to multiclass classification problems, particularly to the multi-label case in which each example may belong to more than one class. We give two boosting methods for this problem. One of these leads to a new method for handling the single-label case which is simpler but as effective as techniques suggested by Freund and Schapire. Finally, we give some experimental results comparing a few of the algorithms discussed in this paper. }, url = {http://www.boosting.org/papers/SchSin99b.pdf}, } @ARTICLE{schapire90strength, author = {Robert E. Schapire}, title = {The Strength of Weak Learnability}, journal = {Machine Learning}, volume = 5, pages = {197-227}, year = 1990, url = {http://citeseer.nj.nec.com/schapire90strength.html}, } @INPROCEEDINGS{schapire99theoretical, author = {Robert E. Schapire}, title = {Theoretical Views of Boosting and Applications}, booktitle = {Algorithmic Learning Theory, 10th International Conference, {ALT} '99, Tokyo, Japan, December 1999, Proceedings}, volume = 1720, publisher = {Springer}, pages = {13--25}, year = 1999, url = {http://citeseer.nj.nec.com/article/schapire99theoretical.html}, } @INPROCEEDINGS{schapire98improved, author = {Robert E. Schapire and Yoram Singer}, title = {Improved Boosting Algorithms using Confidence-Rated Predictions}, booktitle = {Computational Learning Theory}, pages = {80-91}, year = 1998, url = {http://citeseer.nj.nec.com/schapire99improved.html}, } @ARTICLE{schapire99improved, author = {Robert E.~Schapire and Yoram Singer}, title = {Improved Boosting Using Confidence-rated Predictions}, journal = {Machine Learning}, volume = 37, number = 3, year = 1999, pages = {297--336}, } @TECHREPORT{scott98parcel, author = {M. Scott and M. Niranjan and R. Prager}, title = {Parcel: Feature subset selection in variable cost domains}, number = {CUED/F-INFENG/TR 323}, institution = {Cambridge University}, year = 1998, url = {http://citeseer.nj.nec.com/scott98parcel.html}, } @ARTICLE{sharkey97combining, author = {A. Sharkey and N. Sharkey}, title = {Combining diverse neural networks}, journal = {The Knowledge Engineering Review}, year = 1997, volume = 12, pages = {231-247}, number = 3, } @INBOOK{sharkey98:book, publisher = {Springer-Verlag}, pages = {1--30}, year = 1999, author = {Amanda Sharkey}, chapter = {Combining Artificial Neural Nets: Ensemble and Modular Multi-Net Systems}, title = {Multi-Net Systems}, } @INPROCEEDINGS{sharkey97:diversity, booktitle = {Neural Networks and their Applications (NEURAP'97)}, pages = {205--212}, year = 1997 , author = {Amanda Sharkey and Noel Sharkey}, title = {Diversity, Selection, and Ensembles of Artificial Neural Nets}, } @ARTICLE{sharkeychandroth96, title = {Diverse Neural Net Solutions to a Fault Diagnosis Problem}, year = 1996, journal = {Neural Computing and Applications}, volume = 4 , author = {Amanda Sharkey and Noel Sharkey and Gopinath Chandroth}, pages = {218--227}, } @INPROCEEDINGS{sharkey98:adapting, author = {Amanda Sharkey and Noel Sharkey and Simon Cross}, booktitle = {ICANN `98}, publisher = {Springer-Verlag}, title = {Adapting an ensemble approach for the diagnosis of breast cancer}, year = 1998, pages = {281-286}, } @ARTICLE{sharkey97:arm, author = {Noel Sharkey}, title = {Artificial Neural Networks for Coordination and Control: The Portability of Experimental Representations}, journal = {Robotics and Autonomous Systems}, year = 1997, volume = 22, pages = {345-360}, } @ARTICLE{sharkey95weight, author = {Sharkey,N.E. and Neary,J. and Sharkey,A.J.C.}, title = {{S}earching {W}eight {S}pace for {B}ackpropagation {S}olution {T}ypes}, journal = {{C}urrent {T}rends in {C}onnectionism: {P}roceedings of the 1995 {S}wedish {C}onference on {C}onnectionism}, editor = {Niklasson,L.F. and Boden,M.B.}, pages = {103--120}, year = 1995, } @ARTICLE{sierra98global, author = {A. Sierra and C. Cruz}, title = {Global and Local Neural Network Ensembles}, journal = {Pattern Recognition Letters}, volume = 19, number = 8, pages = {651--655}, year = 1998, pdf = {http://dx.doi.org/10.1016/S0167-8655(98)00042-7}, } @ARTICLE{smith92cooperative, number = 2, pages = {127--149}, title = {Searching for Diverse Cooperative Populations with Genetic Algorithms}, volume = 1, year = 1992 , journal = {Evolutionary Computation}, author = {Robert E. Smith and Stephanie Forrest and Alan S. Perelson}, } @ARTICLE{sollichkrogh96:overfitting, author = {P. Sollich and A. Krogh}, title = {Learning with ensembles: How overfitting can be useful}, pages = {190--196}, booktitle = {Advances in Neural Information Processing Systems}, editor = {David S. Touretzky and Michael C. Mozer and Michael E. Hasselmo}, volume = 8, year = 1996, publisher = {The {MIT} Press}, abstract = {We study the characteristics of learning with ensembles. Solving exactly the simple model of an ensemble of linear students, we find surprisingly rich behaviour. For learning in large ensembles, it is advantageous to use under-regularized students, which actually over-fit the training data. Globally optimal performance can be obtained by choosing the training set sizes of the students appropriately. For smaller ensembles, optimization of the ensemble weights can yield significant improvements in ensemble generalization performance, in particular if the individual students are subject to noise in the training process. Choosing students with a wide range of regularization parameters makes this improvement robust against changes in the unknown level of noise in the training data.}, } @ARTICLE{swann98:fastcommittee, volume = 34, number = 14, title = {Fast Committee Learning: Preliminary Results}, month = {July}, journal = {Electronics Letters}, pages = {1408-1410}, year = 1998, author = {A. Swann and N. Allinson}, } @ARTICLE{taniguchi97averaging, author = {Michiaki Taniguchi and Volker Tresp}, title = {Averaging Regularized Estimators}, journal = {Neural Computation}, volume = 9, number = 5, pages = {1163-1178}, year = 1997, } @ARTICLE{tetko95overfitting, author = {Tetko, I. V. and Livingstone, D. J. and Luik, A. I.}, title = {Neural network studies. 1. Comparison of overfitting and overtraining}, journal = {Journal of Chemical Information & Computer Sciences}, volume = 35, number = 5, pages = {826-833}, abstract = {The application of feed forward back propagation artificial neural networks with one hidden layer (ANN) to perform the equivalent of multiple linear regression (MLR) has been examined using artificial structured data sets and real literature data. The predictive ability of the networks has been estimated using a training/test set protocol. The results have shown advantages of ANN over MLR analysis. The ANNs do not require high order terms or indicator variables to establish complex structure-activity relationships. Overfitting does not have any influence on network prediction ability when overtraining is avoided by cross-validation. Application of ANN ensembles has allowed the avoidance of chance correlations and satisfactory predictions of new data have been obtained for a wide range of numbers of neurons in the hidden layer. [References: 30]}, keywords = {Qsar}, year = 1995, } @ARTICLE{tetko93structure, author = {Tetko, I. V. and Luik, A. I. and Poda, G. I.}, title = {Applications of neural networks in structure-activity relationships of a small number of molecules}, journal = {Journal of Medicinal Chemistry}, volume = 36, number = 7, pages = {811-4}, abstract = {We investigated the applications of back propagation artificial neural networks (ANN) for a small dataset analysis in the field of structure-activity relationships. The derivatives of carboquinone were used as an example. It\'s been found that in this case the use of the same neural network results in unambiguous classification of new molecules. Predictions can be improved with statistical analysis of independent prognosis sets. We suggest that the sign criterion be used as a classification rule. We also compared neural networks with FALS and ALS in leave-one-out prediction. ANN applied to the same dataset has shown the same predictive ability as ALS but poorer than FALS.}, keywords = {*Carbazilquinone/aa [Analogs & Derivatives] Comparative Study *Mitomycins/pd [Pharmacology] *Neural Networks (Computer) Structure-Activity Relationship}, year = 1993, } @ARTICLE{tetko97epa, author = {Tetko, I. V. and Villa, A. E.}, title = {An efficient partition of training data set improves speed and accuracy of cascade-correlation algorithm}, journal = {Neural Processing Letters}, volume = 6, number = {1-2}, pages = {51-59}, abstract = {This study extends an application of efficient partition algorithm (EPA) for artificial neural network ensemble trained according to Cascade Correlation Algorithm. We show that EPA allows to decrease the number of cases in learning and validated data sets. The predictive ability of the ensemble calculated using the whole data set is not affected and in some cases it is even improved. It is shown that a distribution of cases selected by this method is proportional to the second derivative of the analyzed function}, keywords = {algorithm, cascade correlation, early stopping, efficient partition of training data set}, year = 1997, } @ARTICLE{tetko97overfitting, author = {Tetko, I. V. and Villa, A. E.}, title = {An enhancement of generalization ability in cascade correlation algorithm by avoidance of overfitting/overtraining problem}, journal = {Neural Processing Letters}, volume = 6, number = {1-2}, pages = {43-50}, abstract = {The current study investigates a method for avoidance of an overfitting/overtraining problem in Artificial Neural Network (ANN) based on a combination of two algorithms: Early Stopping and Ensemble averaging (ESE). We show that ESE provides an improvement of the prediction ability of ANN trained according to Cascade Correlation Algorithm. A simple algorithm to estimate the generalization ability of the method according to the Leave-One-Out technique is proposed and discussed. In the accompanying paper the problem of optimal selection of training cases is considered for accelerated learning of the ESE method}, keywords = {cascade correlation algorithm, early stopping, overfitting, overtraining}, year = 1997, } @ARTICLE{tetko97partition, author = {Tetko, I. V. and Villa, A. E.}, title = {Efficient Partition of Learning Data Sets for Neural Network Training}, journal = {Neural Networks}, volume = 10, number = 8, pages = {1361-1374.}, abstract = {This study investigates the emerging possibilities of combining unsupervised and supervised learning in neural network ensembles. Such strategy is used to get an efficient partition of a noisy input data set in order to focus the training of neural networks on the most complex and informative domains of the data set and accelerate the learning phase. The proposed algorithm provides a good prediction accuracy using fewer cases from non-informative domains according to a correlative measure of dependency between cases of the training set. This measure takes into account internal relationships amid analyzed data and can be used to cluster neighbor cases in a multidimensional space and to filter out the outliers. The possible relation of the proposed algorithm to brain processing occurring in the thalamo-cortical pathway is discussed.}, year = 1997, } @ARTICLE{tetko96variable, author = {Tetko, I. V. and Villa, A. E. and Livingstone, D. J.}, title = {Neural network studies. 2. Variable selection}, journal = {Journal of Chemical Information & Computer Sciences}, volume = 36, number = 4, pages = {794-803}, abstract = {Quantitative structure-activity relationship (QSAR) studies usually require an estimation of the relevance of a very large set of initial variables. Determination of the most important variables allows theoretically a better generalization by all pattern recognition methods. This study introduces and investigates five pruning algorithms designed to estimate the importance of input variables in feed-forward artificial neural network trained by back propagation algorithm (ANN) applications and to prune nonrelevant ones in a statistically reliable way. The analyzed algorithms performed similar variable estimations for simulated data sets, but differences were detected for real QSAR examples. Improvement of ANN prediction ability was shown after the pruning of redundant input variables. The statistical coefficients computed by ANNs for QSAR examples were better than those of multiple linear regression. Restrictions of the proposed algorithms and the potential use of ANNs are discussed.}, keywords = {Databases, Factual Linear Models *Neural Networks (Computer) Nonlinear Dynamics Pharmaceutical Preparations/ch [Chemistry] Structure-Activity Relationship Support, Non-U.S. Gov\'t}, year = 1996, } @ARTICLE{tetko98pruning, author = {Tetko, I. V. and Villa, A. E. P. and Aksenova, T. I. and Zielinski, W. L. and Brower, J. and Collantes, E. R. and Welsh, W. J.}, title = {Application of a pruning algorithm to optimize artificial neural networks for pharmaceutical fingerprinting}, journal = {Journal of Chemical Information & Computer Sciences}, volume = 38, number = 4, pages = {660-668}, abstract = {The present study investigates an application of artificial neural networks (ANNs) for use in pharmaceutical fingerprinting. Several pruning algorithms were applied to decrease the dimension of the input parameter data set. A localized fingerprint region was identified within the original input parameter space from which a subset of input parameters was extracted leading to enhanced ANN performance. The present results confirm that ANNs can provide a fast, accurate, and consistent methodology applicable to pharmaceutical fingerprinting. [References: 26]}, year = 1998, } @BOOK{tikhonov77book, author = {A. N. Tikhonov and V. Y. Arsenin}, title = {Solutions of Ill-posed problems}, publisher = {W.H.Winston and Sons, Washington D.C.}, year = 1977, } @INPROCEEDINGS{ting97stacked, author = {Kai Ming Ting and Ian H. Witten}, title = {Stacked Generalization: When Does It Work?}, booktitle = {{IJCAI} (2)}, pages = {866--873}, year = 1997, url = {http://citeseer.nj.nec.com/ting97stacked.html}, } @INPROCEEDINGS{tresp95combining, author = {Volker Tresp and Michiaki Taniguchi}, title = {Combining Estimators Using Non-Constant Weighting Functions}, pages = {419--426}, editor = {G. Tesauro and D. Touretzky and T. Leen}, volume = 7, booktitle = {Advances in Neural Information Processing Systems}, year = 1995, publisher = {The {MIT} Press}, abstract = {This paper discusses the linearly weighted combination of estimators in which the weighting functions are dependent on the input. We show that the weighting functions can be derived either by evaluating the input dependent variance of each estimator or by estimating how likely it is that a given estimator has seen data in the region of the input space close to the input pattern. The latter solution is closely related to the mixture of experts approach and we show how learning rules for the mixture of experts can be derived from the theory about learning with missing features. The presented approaches are modular since the weighting functions can easily be modified (no retraining) if more estimators are added. Furthermore, it is easy to incorporate estimators which were not derived from data such as expert systems or algorithms.}, } @ARTICLE{tumerghosh94framework, author = {K. Tumer and J. Ghosh}, title = {A framework for estimating performance improvements in hybrid pattern classifiers}, journal = {World Congress on Neural Networks}, volume = 3, pages = {220--5}, publisher = {Lawrence Erlbaum Associates}, month = {June}, year = 1994, abstract = {Classification methods often perform significantly below Bayesian limits in complex, high-dimensional classification tasks because of model bias, inadequate training data and noise/variability in the data. When several classifiers are used for a given task, selecting one method over all others discards potentially valuable information. Strategies aimed at suitably combining the results of multiple classifiers are expected to perform better than any single method, and reduce overall bias and noise. An underwater passive sonar data set consisting of over 1000 samples processed to produce different 25-dimensional and 24-dimensional feature vectors is used in this study to examine an evidence combination framework. An analysis of the conditions that the data sets must satisfy, and the conditions under which improvements can be obtained is provided, and results are presented for hybrid networks using both local and global classifiers. }, } @INPROCEEDINGS{tumerghosh94limits, author = {K. Tumer and J. Ghosh}, title = {Limits to Performance Gains in Combined Neural Classifiers}, editor = {Dagli and Akay and Chen and Fernandez and Ghosh}, booktitle = {Intelligent engineering systems through artificial neural networks: proceedings of the Artificial Neural Networks in Engineering ({ANNIE} '95) Conference}, publisher = {American Society of Mechanical Engineers}, address = {United Engineering Center, 345 E. 47th St., New York, NY 10017, USA}, year = 1994, volume = 5, series = {ASME Press Series on International Advances in Design Productivity}, abstract = {The performance of a single classifier is often inadequate in difficult classification problems. In such cases, several researchers have combined the outputs of multiple classifiers to obtain better performance. However, the amount of improvement possible through such combination techniques is generally not known. We present two approaches to estimating performance limits in hybrid networks. First, we present a framework that estimates Bayes error rates when linear combiners are used. Then we discuss a more general method that provides decision confidences and error bounds based on error types arising from the training data. The methods are illustrated for a difficult four class problem involving underwater acoustic data. For this data, we compute the single classifier and combiner classification performances, as well as the Bayes error rate and an error bound.}, } @INPROCEEDINGS{tumerghosh95boundary, author = {K. Tumer and J. Ghosh}, title = {Boundary variance reduction for improved classification through hybrid networks}, booktitle = {Proceedings of the Spie Conf. on Applications and Science of Artificial Neural Networks IV}, volume = 2492, pages = {573--585}, month = {April}, year = 1995, address = {Orlando, FL}, abstract = {Several researchers have shown experimentally that substantial improvements can be obtained in difficult pattern recognition problems by combining or integrating the outputs of multiple classifiers. This paper provides an analytical framework that quantifies the improvements in classification results due to linear combination. We show that combining networks in the output space reduces the variance of the actual decision region boundaries around the optimum boundary. In the absence of network bias, the added classification error is directly proportional to the boundary variance. Moreover, if the network errors are independent, then the reduction in variance boundary location is by a factor of N, the number of classifiers that are combined. In the presence of network bias, the reductions are less than or equal to N, depending on the interaction between network biases. We discuss how the individual networks can be selected to achieve significant gains through combination, and we support them with experimental results on 25-dimensional sonar data. The analysis presented facilitates the understanding of the relationships among error rates, classifier boundary distributions and combination in the output space.}, } @ARTICLE{tumerghosh95orderstats, author = {K. Tumer and J. Ghosh}, title = {Order Statistics Combiners for Neural Classifiers}, journal = {World Congress on Neural Networks}, volume = {Vol. I}, pages = {31--34}, publisher = {INNS Press}, address = {Washington, DC}, month = {July}, year = 1995, abstract = {Several researchers have shown that linearly combining outputs of multiple neural classifiers results in better performance for many applications. In this paper we introduce a family of order statistics combiners as an alternative to linear combiners. We show analytically that the selection of the median, the maximum and in general, the i-th order statistic improves classification performance. Specifically, we show that order statistics combiners reduce the variance of the actual decision boundaries around the optimum boundary, and that this is directly related to classification error.}, } @TECHREPORT{tumerghosh95theoretical, author = {Kagan Tumer and Joydeep Ghosh}, title = {Theoretical Foundations of Linear and Order Statistics Combiners for Neural Pattern Classifiers}, institution = {Computer and Vision Research Center, University of Texas, Austin}, year = 1995, number = {TR-95-02-98}, url = {http://citeseer.nj.nec.com/tumer96theoretical.html}, } @ARTICLE{tumerghosh96, author = {Kagan Tumer and Joydeep Ghosh}, title = {Error Correlation and Error Reduction in Ensemble Classifiers}, journal = {Connection Science}, volume = 8, number = {3-4}, pages = {385--403}, year = 1996, abstract = {Using an ensemble of classifiers, instead of a single classifier, can lead to improved generalization. The gains obtained by combining, however, are often affected more by the selection of what is presented to the combiner than by the actual combining method that is chosen. In this paper, we focus on data selection and classifier training methods, in order to 'prepare' classifiers for combining. We review a combining framework for classification problems that quantifies the need for reducing the correlation among individual classifiers. Then, we discuss several methods that make the classifiers in an ensemble more complementary. Experimental results are provided to illustrate the benefits and pitfalls of reducing the correlation among classifiers, especially when the training data are in limited supply.}, } @ARTICLE{tumerghosh96analysis, author = {Kagan Tumer and Joydeep Ghosh}, title = {Analysis of decision boundaries in linearly combined neural classifiers}, journal = {Pattern Recognition}, year = 1996, volume = 29, number = 2, pages = {341--348}, month = {February}, } @INPROCEEDINGS{tumerghosh96estimating, author = {Kagan Tumer and Joydeep Ghosh}, title = {Estimating the Bayes Error Rate Through Classifier Combining}, booktitle = {Proceedings of the 13th International Conference on Pattern Recognition}, month = {August}, year = 1996, abstract = {The Bayes error provides the lowest achievable error rate for a given pattern classification problem. There are several classical approaches for estimating or finding bounds for the Bayes error. One type of approach focuses on obtaining analytical bounds, which are both difficult to calculate and dependent on distribution parameters that may not be known. Another strategy is to estimate the class densities through non-parametric methods, and use these estimates to obtain bounds on the Bayes error. This article presents a novel approach to estimating the Bayes error based on classifier combining techniques. For an artificial data set where the Bayes error is known, the combiner-based estimate outperforms the classical methods.}, } @ARTICLE{tumer98rbfmedical, author = {Kagan Tumer and Nirmala Ramanujam and Joydeep Ghosh and Rebecca Richards-Kortum}, title = {Ensembles of Radial Basis Function Networks for Spectroscopic Detection of Cervical Pre-Cancer}, journal = {IEEE Transactions on Biomedical Engineering}, year = 1998, volume = 45, number = 8, pages = {953--961}, abstract = {Medical applications usually used Radial Basis Function Networks just as Artificial Neural Networks. However, RBFNs are Knowledge-Based Networks that can be interpreted in several way: Artificial Neural Networks, Regularization Networks, Support Vector Machines, Wavelet Networks, Fuzzy Controllers, Kernel Estimators, Instanced-Based Learners. A survey of their interpretations and of their corresponding learning algorithms is provided as well as a brief survey on dynamic learning algorithms. RBFNs' interpretations can suggest applications that are particularly interesting in medical domains. }, } @INPROCEEDINGS{uedanakano96, booktitle = {Proceedings of International Conference on Neural Networks}, title = {Generalization Error of Ensemble Estimators}, author = {N. Ueda and R. Nakano}, year = 1996 , pages = {90--95}, annote = {First presentation (in ML literature) of the bias-variance-covariance decomposition.} } @INPROCEEDINGS{wahba99biasvariance, author = {G. Wahba and X. Lin and F. Gao and D. Xiang and R. Klein and B. Klein}, title = {The bias-variance tradeoff and the randomized {GACV}}, booktitle = {Advances in Neural Information Processing Systems}, number = 11, pages = {620--626}, publisher = {MIT Press}, editor = {M. Kearns and S. Solla and D. Cohn}, year = 1999, url = {http://citeseer.nj.nec.com/wahba99biasvariance.html}, } @ARTICLE{wan90bayesneural, author = {E. A. Wan}, title = {Neural network classification: A Bayesian interpretation}, journal = {IEEE Transactions on Neural Networks}, type = {Letter}, year = 1990, volume = 1, number = 4, pages = {303--304}, } @INPROCEEDINGS{wan97:sunspots, author = {Eric A. Wan}, booktitle = {International Conference On Neural Networks (ICNN97)}, year = 1997 , title = {Combining Fossil and Sunspot Data: Committee Predictions}, url = {http://citeseer.ist.psu.edu/146595.html}, } @INPROCEEDINGS{wezel98maximum, author = {M.C. van Wezel and W.A. Kosters and J.N. Kok}, title = {Maximum Likelihood Weights for a Linear Ensemble of Regression Neural Networks}, booktitle = {Proceedings International Conference in Neural Information Processing (ICONIP\'98)}, year = 1998, publisher = {IOS Press}, editor = {S. Usui and T. Omori}, pages = {498--501}, address = {Kitakyushu, Japan}, } @ARTICLE{windeatt97spectral, author = {Windeatt, T. and Tebbs, R.}, title = {Spectral technique for hidden layer neural network training}, journal = {Pattern Recognition Letters}, volume = 18, issue = 8, pages = 723, year = 1997, month = {August}, } @ARTICLE{wolpert92stacked, author = {D. H. Wolpert}, title = {Stacked Generalization}, journal = {Neural Networks}, volume = 5, year = 1992, pages = {241--259}, abstract = {This paper introduces stacked generalization, a scheme for minimizing the generalization error rate of one or more generalizers. Stacked generalization works by deducing the biases of the generalizer(s) with respect to a provided learning set. This deduction proceeds by generalizing in a second space whose inputs are (for example) the guesses of the original generalizers when taught with part of the learning set and trying to guess the rest of it, and whose output is (for example) the correct guess. When used with multiple generalizers, stacked generalization can be seen as a more sophisticated version of cross-validation, exploiting a strategy more sophisticated than cross-validation's crude winner-takes-all for combining the individual generalizers. When used with a single generalizer, stacked generalization is a scheme for estimating (and then correcting for) the error of a generalizer which has been trained on a particular learning set and then asked a particular question. After introducing stacked generalization and justifying its use, this paper presents two numerical experiments. The first demonstrates how stacked generalization improves upon a set of separate generalizers for the NETtalk task or translating text to phonemes. The second demonstrates how stacked generalization improves the performance of a single surface-fitter. With the other experimental evidence in the literature, the usual arguments supporting cross-validation, and the abstract justifications presented in this paper, the conclusion is that for almost any real-world generalization problem one should use some version of stacked generalization to minimize the generalization error rate. This paper ends by discussing some of the variation of stacked generalization, and how it touches on other fields like chaos theory. }, } @ARTICLE{wolpert97bias, author = {David Wolpert}, title = {On Bias Plus Variance}, journal = {Neural Computation}, volume = 9, number = 6, pages = {1211-1243}, year = 1997, url = {http://citeseer.nj.nec.com/article/wolpert96bias.html}, } @ARTICLE{woods97local, author = {K. Woods and W.P. Kegelmeyer and K. Bowyer}, title = {Combination of multiple classifiers using local accuracy estimates}, journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence}, year = 1997, volume = 19, pages = {405-410}, } @INCOLLECTION{yao99review, publisher = {IEEE}, pages = {1423-1447}, volume = 87 , year = 1999, author = {Xin Yao}, number = 9, title = {Evolving Artificial Neural Networks}, month = {September}, booktitle = {Proceedings of the IEEE}, } @INCOLLECTION{yaoliu96, author = {Xin Yao and Yong Liu}, pages = {229-242}, booktitle = {Complex Systems - From Local Interactions to Global Phenomena}, publisher = {IOS Press, Amsterdam}, title = {How to Make Best Use of Evolutionary Learning}, year = 1996, } @ARTICLE{yaoliu97epnet, author = {Xin Yao and Yong Liu}, number = 3, pages = {694--713}, journal = {IEEE Transactions on Neural Networks}, year = 1997, title = {A new evolutionary system for evolving artificial neural networks}, month = {May}, volume = 8, } @INCOLLECTION{yaoliu98making_use, pages = {417--425}, booktitle = {IEEE Transactions on Systems, Man and Cybernetics, Part B: Cybernetics}, year = 1998, volume = 28, number = 3, publisher = {IEEE Press}, title = {Making use of Population Information in Evolutionary Artificial Neural Networks}, month = {June}, author = {Xin Yao and Yong Liu}, } @INPROCEEDINGS{yaoliu99breastcancer, author = {Xin Yao and Yong Liu}, title = {Neural networks for breast cancer diagnosis}, year = 1999, booktitle = {Proceedings of the 1999 Congress on Evolutionary Computation}, pages = {1760-1767}, volume = 3, month = {July}, publisher = {IEEE Press}, } @ARTICLE{yates95use, author = {W. Yates and D. Partridge}, title = {Use of methodological diversity to improve neural network generalization}, journal = {Neural Computing and Applications}, year = 1996, volume = 4, pages = {114--128}, number = 2, url = {http://citeseer.nj.nec.com/partridge95use.html}, } @ARTICLE{condorcet88review, author = {P. Young}, title = {Condorcet's Theory of Voting}, journal = {American Political Science Review}, year = 1988, volume = 82, number = {1231-1244}, }