Parallel generating RDFS closure for reasoning

The closure of an RDF(S) source reveals the implicit semantics that it implies. In order to improve reasoning efficiency, researchers proposed to generate the RDFS closure for reasoning on RDF(S) datasets. Serially generating closure is insufficient to large scale RDF(S) source because of too much time taking and memory demand. In this paper, we present an approach of parallel generating the RDFS closure. We analyze the RDFS entailment rules, classify RDF(S) triples of the source according to the forms of triples appearing in the rules and partition RDFS source into several subsets. Furthermore, we discuss parallel mechanism of RDF(S) source and distribute the subsets to parallel process for generating closure by applying the RDFS entailment rules. The results of analysis and experiments show that parallel approach increase the efficiency of reasoning. Our approach is suitable for reasoning over large scale RDF(S) source.