Normalized Medline Distance and Its Utilization in Context-aware Life Science Literature Search

When facing great volume of query results while users are searching literatures on the Web, we propose to refine the search process by using user interests. We analyze user interests and calculate semantic similarity among those interest terms to fulfill query refinement. Traditional general purpose similarity measures may not always fit a domain specific context. In this paper, under the context of medical literature search on the Web, based on the biomedical literature knowledge source “Medline”, we propose a novel similarity method—Normalized Medline Distance, which could deal with some problems of existing similarity measures and reflect relevance between medical terms more reasonably. Based on the proposed measure method, more accurate user interest description could be acquired. Based on the ranked user interests, when users try to search literatures by keywords, the most relevant results which are related to their current interests and the query input would be ranked to the top.