Related Experiment Videos
Context-based ontology building support in clinical domains using formal concept analysis
Guoqian Jiang1, Katsuhiko Ogasawara, Akira Endoh
1Department of Medical Informatics, Hokkaido University Graduate School of Medicine, North 15, West 7, Kita-ku, Sapporo 060-8638, Japan. guoqian@med.hokudai.ac.jp
International Journal of Medical Informatics
|August 12, 2003
Summary
Formal Concept Analysis (FCA) aids clinical ontology building by extracting meaningful medical concepts from discharge summaries. This system supports experts in leveraging linguistic and contextual knowledge for improved medical informatics.
Area of Science:
- Medical Informatics
- Ontology Engineering
- Clinical Terminology
Background:
- Clinical domain ontology development is crucial in medical informatics.
- Existing methods for building clinical ontologies require enhancement.
- Formal Concept Analysis (FCA) offers a potential approach for structured knowledge acquisition.
Purpose of the Study:
- To explore the utility of Formal Concept Analysis (FCA) for supporting context-based ontology construction in clinical domains.
- To investigate the application of FCA in cardiovascular medicine for ontology development.
- To evaluate an integrated system combining FCA and Natural Language Processing (NLP) for clinical ontology support.
Main Methods:
- Development of an ontology building support system integrating FCA and NLP modules.
- Implementation as a Protégé-2000 JAVA tab plug-in.
- Utilized 368 Japanese discharge summaries and the MEDIS ver2.0 diagnostic terms dictionary as knowledge sources.
Main Results:
- Achieved stable and high-precision medical concept extraction from the MEDIS dictionary.
- 73+/-14% of extracted compound medical phrases were clinically meaningful concepts.
- 57.7% of extracted attribute implication pairs (medical concept pairs) were clinically validated.
Conclusions:
- The developed ontology building support system effectively leverages FCA for clinical knowledge extraction.
- Clinical experts can utilize the system to access linguistic information and context-based knowledge.
- The system demonstrates utility in supporting clinical ontology building tasks.