USING LATENT SEMANTIC INDEXING FOR LITERATURE-BASED DISCOVERY

Latent semantic indexing (LSI) is a statistical technique for improvin g information retrieval effectiveness. Here, we use LSI to assist in l iterature-based discoveries. The idea behind literature-based discover ies is that different authors have already published certain underlyin g scientific ideas that, when taken together, can be connected to hypo thesize a new discovery, and that these connections can be made by exp loring the scientific literature. We explore latent semantic indexing' s effectiveness on two discovery processes: uncovering ''nearby'' rela tionships that are necessary to initiate the literature based discover y process; and discovering more distant relationships that may genuine ly generate new discovery hypotheses.