Semantic based real-time clustering for PubMed literatures

This paper addresses to use the latent semantic topology to real-time cluster the literatures retrieved by PubMed in response to clinical queries and evaluates its performance by professional experts. The result shows that semantic clusters properly offer an exploratory view on the returned search results, which saves users' time to understand them. Besides, most experts conceive that the documents assigned to the identical cluster are similar and the concepts of clusters are appropriate.