Information retrieval techniques for knowledge discovery in biomedical literature

This paper presents our contribution to enhance literature-based discovery with information retrieval techniques. We propose the joint use of a flexible Information Retrieval model and MeSH Concepts for knowledge discovery in biomedical literature. The Information Retrieval model contributes to filter MEDLINE biomedical literature to the most relevant documents. Utilizing MeSH concepts allows to quickly identifying candidate concepts that could potentially validate a hypothesis. We have tested our approach by replicating the Swanson's first discovery on fish oil and Raynaud's disease correlation. The obtained results show the effectiveness of our approach.