fp-Sparql: An RDF fuzzy retrieval mechanism supporting user preference

RDF fuzzy retrieval is an important module for realizing intelligent retrieval of Web semantics. In this paper, Zadeh's type-II fuzzy set theory, as well as the concepts of α-cut set and linguistic variable is adopted to put forward the RDF fuzzy retrieval mechanism supporting user preference, which extends SPARQL language for further realizing fuzzy and preferred expression. Moreover, ordered sub-domain table of linguistic values is constructed, to realize the projection from the fuzzy values to relayed sub-domains in the table, so as to figure out the interval of membership. On this basis, it will then be converted into the standard SPARQL by drawing support from the defuzzification rules, so as to achieve fuzzy retrieval operations. In order to test the ideology proposed in this paper, the fp-SPARQL (fuzzy and preference SPARQL) retrieval system is developed. The system has extended univ_Bench main body, with plenty of retrieval instances constructed. On this basis, visual methods will be employed to realize a fuzzy and preferred retrieval on RDF. According to the result of this experiment, our method has improved the performance of RDF fuzzy retrieval, and correspondingly, users' satisfaction rate on the retrieval results is also enhanced. Besides, the retrieval method complies with users' habit, endowing the system with higher flexibility nature.