GIRS : genetic information retrieval system.

Abstract

Currently, companies and corporations deal with a big amount of internal documents (rules, contracts, resolutions, records, communications, others) generated daily. Such documents are stored in several places and formats and, consequently, there is a need to restore information found in these documents. The problem is that the quantity and diversity of documents stored make it difficult for users to find useful information in the files. Moreover, there is a great difficulty in finding truly relevant information that is often lost among a great number of less important documents. This thesis brings a proposal of an information retrieval system where the value of restored documents increases in each interaction through genetic algorithms with direct help of the system users and trough an implicit feedback, aiming at the improvement of restored documents’ relevance every new search.