AT&T at TREC-7

This year AT&T participated in the ad-hoc task and the Filtering, SDR, and VLC tracks. Most of our e ort for TREC-7 was concentrated on SDR and VLC tracks. On the ltering track, we tested a preliminary version of a text classi cation toolkit that we have been developing over the last year. In the ad-hoc task, we introduce a new tf-factor in our term weighting scheme and use a simpli ed retrieval algorithm. The same weighting scheme and algorithm are used in the SDR and the VLC tracks. The results from the SDR track show that retrieval from automatic transcriptions of speech is quite competitive with doing retrieval from human transcriptions. Our experiments indicate that document expansion can be used to further improve retrieval from automatic transcripts. Results of ltering track are in line with our expectations given the early developmental stage of our classi cation software. The results of VLC track do not support our hypothesis that retrieval lists from a distributed search can be e ectively merged using only the initial part of the documents.