Research On RDF Ontology Querying Using Query Relaxation Based Similarity And Contextual Preferences

The Semantic Web is a new vision of the next generation of the Web and the development direction and an extension of the current Internet, whose goal is to make network application more intelligent and automated and to make the information on the Web can be understood by machines and to make people retrieve information more efficiently. As carriers of knowledge in the Semantic Web, RDF (Resource Description Framework) makes the semantic retrieval in the the Semantic Web to be possible. With the size and complexity of RDF ontology increasing rapidly, more and more query demand on the RDF ontology exists and becomes increasingly difficult.In the RDF ontology query, on the one hand, the structure and contents of heterogeneous ontology schema and complex query languages are always not known to users. Even if users have clear intentions, however, they also aren’t able to correctly formulate queries to databases which will not return any answer or a small number of results in response to a user query. In this situation, relaxing the original query for presenting more relevant items is desired to the users. On the other hand, after the query relaxation, however, users may be confronted with sevrel relaxing paths and many answers problem. There will be a great semantic distant between some of the results and the initial query, and then users will want the system to have the option of ordering the matches automatically in order to deal with information overload.Aiming at the above problems, this paper presents a method using query relaxation based similarity and contextual preferences. To deal with the problem of empty answers or a small number of results, a RDF query relaxation method is propsed which through using the rules of RDFS and domain preferences to rewrite the initial query model to get more results. During the relaxation process, the relaxed query tree model is constructed and based on different paths the relaxation will provide more results. To deal with the problem of multi-relaxing strategies and many answers returned in response to a relaxed query, a RDF semantic similarity and contextual preferences-based modle is proposed. Based on the quantitative representation of the semantic similarity of relaxed query modle and the contextual preference degree of rewriting rules, the relaxing paths are weighted to provide many relaxing path and answers ranking approach. When the query comes, these orders, with the semantic similarity which is between the initial query and the relaxed query and the contextual preferences of rewriting rules, are used to expeditiously provide top-k ranked result tuples. On the basis of the above query relaxation strategy, the paper presents a query relaxation algorithm and finally realizes a simple prototype system and through which we verify that the proposed query relaxation method is feasible. Besides, we also make some comparisons with the existing RDF query relaxation methods and prove that our method has superiority in response time and recall precision.