Research Of RDF Query Based On Inference Mechanism

With the rapid development of the semantic Web, the information in the semantic Web grows explosively, and how to obtain the useful information quickly and accurately becomes a hot topic. Because RDF (Resource Description Framework) is a general standard of metadata model in the semantic Web, RDF data is being used more and more widely. Therefore, the development of RDF query technology affects the development of the semantic Web directly.Currently, there are many kinds of RDF query technologies, and the SPARQL, a RDF query language, has been identified as the recommended standard language by the W3C (World Wide Web Consortium). But a series of problems still exist in the current RDF query technology. On the one hand, because of the concealment of the RDF information in the semantic Web, the RDF query cannot mine the RDF information deeply just by the traditional RDF query technology. On the other hand, ordinary users are different from the relevant researchers who understand the knowledge of RDF query and the inner structure of the RDF information, so when the ordinary users do the query, the condition of the query is very fuzzy and uncertain. All of these factors lead to a series of problems about query, such as too few query results, and even empty result. In order to solve these problems, this thesis combines the inference mechanism and RDF query technology together to propose a RDF query method based on inference mechanism.Based on the traditional RDFS (RDF Schema) inference rules, this thesis designs and implements a RDFS inference algorithm used in the RDF query, and adopts a sorting algorithm to process the results after the query. Firstly, it analyzes the RDFS contain rules thoroughly at the beginning of this thesis, and according to the related triples and the relationship between attributes in the RDFS contain rules, this thesis optimizes the RDFS rules and obtains an optimal sequence of execution. Based on the optimal sequence of rules, we can design the optimized inference algorithm. Secondly, this thesis introduces the concept of semantic distance into the rewrite of RDF query and this method can reduce the complexity of time and space by optimizing the conditions of query rewrite. Thirdly, after the RDF query, there are a vast amount of RDF triples, and how to select the closest information to the users’ demand is a big and complex problem. This thesis adopts the Top-k sorting algorithm based on the TA algorithm, and this sorting method combines the concept of closeness to calculate the comprehensive score, so this algorithm can return the information that can meet the users’ requirements uttermost quickly. Finally, this thesis designs the experiment to verify this RDF query method and analyzes the results of the experiment. The results of this experiment show that the RDF query method based on inference mechanism proposed by this thesis is feasible and this query method has a good performance on the recall ratio and precision ratio.