ALCOMFT-TR-03-185

Luigi Laura, Umberto Nanni and Fabiano Sarracco
Query Transformation through approximated LSI computation
Rome. Work package 1. December 2003.

Abstract: The Latent Semantic Indexing (LSI) proved to be an effective technique in the field of Information Retrieval. Its drawbacks are the time needed to compute the SVD decomposition and to answer queries, since the query must be compared against each document in the collection. In this paper we present a technique that uses the information computed by the traditional LSI to provide a fast online answer to the users. Informally, if compared to the traditional LSI, we can say that instead of returning the documents that are related to the concepts in the query, we return the documents that contain terms that are related to the concepts in your query.

Postscript file: ALCOMFT-TR-03-185.ps.gz (289 kb).