An Information Retrieval System with Weighted Querying Based on Multi-Granular Linguistic Information

In this contribution, an information retrieval system (IRS) based on fuzzy multi-granular linguistic information is proposed. The user queries and IRS responses are modelled using different linguistic domains or label sets with different cardinalities and/or semantics. We present a method to process the multigranular linguistic information in the retrieval activity of the IRS. The system accepts Boolean queries whose terms can be simultaneously weighted by means of ordinal linguistic values according to two semantics: a symmetrical threshold semantics and an importance semantics. In both semantics the linguistic weights are represented by the linguistic variable ”Importance”, but assessed on different label sets S1 and S2, respectively. The IRS evaluates the weighted queries and obtains the linguistic retrieval status values (RSV) of documents represented by a linguistic variable ”Relevance” expressed also on a different label set S’. The advantage of this linguistic IRS with respect to others is that the use of the multi-granular linguistic information facilitates the expression of information needs and improves the latter issue.