Research On The Large-scale Ontology Partitioning And Mapping For The Semantic Service

With the rapid development of the Internet, the information and data on the Internet presents an explosive growth, and also to promote the development of the Semantic Web, which make the scale of ontology used to describe the knowledge in different areas becomes more and more larger, it is named as large ontology. Partitioning and mapping the large ontology is one key problem in the research of the semantic web, which is an important means to solve the heterogeneous of ontology and knowledge sharing, it is an important way to promote the semantic technology from the theretical to the pratical application, so this paper focus on the study of partitioning and mapping large ontology.Firstly, this paper describes the research background of the subject, including that the introduction of the semantic web and the relation work of partitioning and mapping large ontology.Secondly, aiming at the problems of the blank blocks can not be matched, the low efficiency of partitioning and mapping, this paper presents an ontology mapping method based on the synchronous partition of the anchors. In this method, the synchronous partition of the anchors is achieved by the strutural similarity, and then the remaining non-anchors of the ontology are assigned based on the computation of the nearest neighbor anchors, finally, a hierarchical diffusion-oriented optimization iterative algorithm is designed to mapping blocks.Thirdly, in order to further ensure the consistency of the mapping, based on block mapping, conflict detection and mapping correction algorithm is presented based on semantic distance and reasoning combined. The algorithm detects the semantic conflict in all mapping sets based on conflict function, and then obtains the mapping conflict sets, which will be repaired by customizing inference rules, so all possible semantic conflicts will be removed to ensure the accuracy the final mapping.Finally, in order to verify the superiority of the proposed method, this paper designs two experiments to evalute the quality of partitioning and mapping by using the the OAEI standard test data sets and Russia12and TourismAB of test data sets. The first experiment assess the quality of the partitioning, the second experiment evaluates the efficiency of dealing with large-scale ontology mapping. The experimental results show that this method can better solve the blank block mismatch problem, and the mismatch between blocks problem. It also can improve the recall and precision of the mapping.