An effective method of large scale ontology matching

Area of Science:

• Biomedical Informatics
• Computational Biology
• Knowledge Representation

Background:

• Increasing heterogeneity of biomedical data sources and knowledge organization systems.
• Challenges in interoperability due to diverse ontologies.
• Limited tools for matching large-scale life science ontologies.

Purpose of the Study:

• To develop an effective method for large-scale ontology matching.
• To address the interoperability gap in biomedical data integration.
• To improve the matching of large ontologies in the life sciences domain.

Main Methods:

• Information Retrieval (IR) techniques.
• Combination of lexical and machine learning for contextual similarity.
• Leveraging synonym terms from the Unified Medical Language System (UMLS).

Main Results:

• Implementation of ServOMap, a fast and efficient large-scale ontology matching system.
• Demonstrated high precision in matching ontologies with hundreds of thousands of entities.
• Ranked among top systems in the Ontology Alignment Evaluation Initiative (OAEI) campaigns (2012, 2013).

Conclusions:

• Proposed approach effectively handles large-scale ontology matching using IR and hybrid similarity measures.
• ServOMap is particularly suited for the life sciences domain due to its use of lexical resources like UMLS.
• The system achieves efficient computation times for large ontologies, facilitating biomedical data interoperability.