Ontologia como recurso de padronização terminológica no processo de recuperação de informação

Abstract

Information Retrieval process involves a collection of documents that must be represented by linguistic expressions and users trying to describe linguistically their information need in order to obtain relevant documents that meet such needs. Therefore, an information retrieval system is a linguistic environment mediating the communication between a stock of information and its users. Its effectiveness depends on adequate control of language for representation of information items and requests of its users. This paper presents a method of using ontologies in information retrieval process. It uses the Vector Space Model as a formal structure for the representation of documents and queries. The documents vectors are created during the automatic indexing process, in which the ontologies provide new terms in order to enrich those representations. The search vector is created from a query expansion process in which, from inferences in ontology, new terms are entered in the search expression initially formulated by the user. The aim is standardize the representations of documents and searches, thus improving the information retrieval process.