A Query Refinement to Concept-based Information Retrieval

This paper presents an approach to semantic Information Retrieval, based on the use of a fuzzy conceptual structure, called Object-Fuzzy Concept Network (O-FCN). It consists of a set of objects ODB stored in a database and of a Fuzzy Concept Network (FCN), which is a complete weighted graph, with edges among concept nodes, we called correlations. It was already introduced and described an Information Retrieval Algorithm involving an OFCN. The algorithm is validated in a classical, crisp, case versus a fuzzy case. The users had the possibility to assign weights of importance to concepts modifiers in queries in order for the system to produce more relevant documents. For a better evaluation it has been finally introduced a new fuzzy accuracy measure that allowed to improve the results.