Comparison and Abstraction of SNOMED CT Ontologies


IAA 228 EPSRC Project
Principal Investigator:
Dr Renate Schmidt
Researcher:
Dr Jieying Chen
Collaborators:
Dr Yongsheng Gao, Dr Dirk Walther, Ghadah Alghamdi, Ian Hyland
Industry partner:
IHTSDO (SNOMED International)
Start Date:
01/09/2018
End Date:
28/02/2019

Summary

Ontologies are widely employed to give structured representations of knowledge suitable for AI reasoning and have become increasingly popular across numerous industry sectors, including medical, digital biology and energy sectors. The project aimed to apply knowledge and tools developed from previous EPSRC-funded research for tracking the changes between different versions of an ontology and forgetting some terms from an ontology thereby computing an abstraction of it. The focus of the project was a trial on a core subset of the SNOMED CT medical ontology, which were subject to an internal quality improvement project at IHTSDO.

Achievements and outcomes

The achievements and outcomes of the project were:

  1. The proposal of a workflow for the computation of ontology extracts of the SNOMED CT ontology of mmedical terms. The workflow incorporates ontology modularity tools, forgetting tools and newly developed signature adjustment techniques.
  2. Development of prototype tool of the workflow, inlcuding a signature adjustment tool implementing different signature extension algorithms that were explored.
  3. Identification of data sets suitably alieniated to actual company/NHS scenarios (European Renal Association, Medical Imaging and NHS refsets).
  4. Extensive evaluation of the workflow on this data using three different modularisation tools, and three different forgetting tools including our own FAME and LETHE.
  5. Evaluation and better understanding of behaviour (ito. success rates, precision and performance) of modularisation and forgetting tools, used in isolation and combination on (fragments of) the SNOMED CT ontology.
  6. Follow-on funding was secured. This new project, a 12 month Secondment, is included in the strategic roadmap for innovation of IHTSDO.
Results and findings of the work were presented at the annual SNOMED Expo event, two international workshops and one international conference.

Data and Software

Bibliography

1
J. Chen, G. Alghamdi, R. A. Schmidt, D. Walther, and Y. Gao.
Modularity meets forgetting: A case study with the SNOMED CT ontology.
In M. Simkus and G. E. Weddell, editors, Proceedings of the 32nd International Workshop on Description Logics (DL'19), volume 2373 of CEUR Workshop Proceedings. CEUR-WS.org, 2019.
2
J. Chen, G. Alghamdi, R. A. Schmidt, D. Walther, and Y. Gao.
Modularity meets forgetting: A case with the SNOMED CT ontology, 2019.
Presented at SNOMED CT Expo 2019.
3
J. Chen, G. Alghamdi, R. A. Schmidt, D. Walther, and Y. Gao.
Modularity meets forgetting: A case with the SNOMED CT ontology (Extended Abstract), 2019.
Presented at PxTP 2019.
4
J. Chen, G. Alghamdi, R. A. Schmidt, D. Walther, and Y. Gao.
Ontology extraction for large ontologies via modularity and forgetting.
In M. Kejriwal, P. A. Szekely, and R. Troncy, editors, Proceedings of the 10th International Conference on Knowledge Capture (K-CAP’19), pages 45–52. ACM, 2019.
5
I. Hyland and R. A. Schmidt.
Protege-TS: An OWL ontology term selection tool.
In Stefan Borgwardt and Thomas Meyer, editors, Proceedings of the 33rd International Workshop on Description Logics (DL 2020), volume 2663 of CEUR Workshop Proceedings. CEUR-WS.org, 2020.