a fuzzy set approach to topical information retrieval.pdfVIP

EUSFLAT - LFA 2005 A fuzzy set approach to concept-based information retrieval Mustapha Baziz * Mohand Boughanem* Gabriella Pasi** Henri Prade* * IRIT, 118 Route de Narbonne, 31062 Toulouse cedex 04 – France, {baziz, boughanem, prade}@irit.fr ** ITC-CNR, Via Bassini 15, 20133 Milano, Italy, gabriella.pasi@r.it Abstract describe the topical content and structure of documents. These efforts gave raise to concept- In this paper an information retrieval based Information Retrieval, which aims at approach is proposed based on the use of a retrieving relevant documents on the basis of their fuzzy conceptual structure used both to meaning rather than their keywords. The main idea index document and to express user queries. at the basis of conceptual IR is that the meaning of a The conceptual structure is hierarchical and text depends on conceptual relationships to objects it encodes the knowledge of the topical in the world rather than to linguistic relations found domain of the considered documents. It is in text or dictionaries [7]. To this aim, sets of words, formally represented as a weighted tree. phrases, names are related to the concepts they The evaluation of conjunctive queries is encode. based on the comparison of minimal sub- In this paper, a fuzzy set approach to concept-based trees