ABSTRACT:

Search queries on biomedical databases, like PubMed, usually come back a giant range of results, solely a small subset of that has relevancy to the user. Ranking and categorization, that will additionally be combined, are proposed to alleviate this data overload downside. Results categorization for biomedical databases is the main target of this work. A natural manner to prepare biomedical citations is per their MeSH annotations. MeSH may be a comprehensive concept hierarchy utilized by PubMed. In this paper, we have a tendency to gift the BioNav system, a completely distinctive search interface that permits the user to navigate massive vary of question results by organizing them using the MeSH concept hierarchy. First, the question results are organized into a navigation tree. At every node growth step, BioNav reveals solely a little subset of the concept nodes, selected such that the expected user navigation price is minimized. In distinction, previous works expand the hierarchy throughout a predefined static manner, while not navigation value modeling. We show that the matter of selecting the only ideas to reveal at each node growth is NP-complete and propose an economical heuristic conjointly a possible optimal algorithm for comparatively very little trees. We have a tendency to tend to point out experimentally that BioNav outperforms state-of-the-art categorization systems by up to an order of magnitude, with respect to the user navigation value.