Combining Evidence For Effective Information Filtering

Susan T. Dumais

As part of NIST/ARPA’s TREC Wcckshop, we used Latent Semantic Indexing (LSl) for filtering 336k incoming documents from diverse sources (newswires, patents, technical abstracts) for 50 topics of interest. We developed representations of user interests or filters for these topics using two sources of training information.


This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.