A system for information retrieval in a medical digital library based on modular ontologies and query reformulation

Ontologies have proven to be useful in the area of Information Retrieval and the biomedical informatics community has acknowledged, in recent years, their utility. However, building and updating manually ontologies is a long and tedious task. This paper proposes a system that allows any search engine to develop its semantic layer by applying ontology learning techniques on Web snippets and applies it to a well-known medical digital library, PubMed. The new system (SemPubMed) automatically builds new ontology fragments related to the user’s query and then it reformulates queries using the new concepts in order to improve information retrieval. Our system has endured a twofold evaluations. On the one hand, we have evaluated the quality of the modular ontologies built by the system. On the other hand, we have studied how the semantic reformulation of the queries has led to an improvement of the quality of the results given by PubMed, both in terms of precision and recall. Obtained results show that adding semantic layer to PubMed enables an improvement of query reformulation and predicted ranking score.