Research On Efficient Multi-Strategy Ontology Mapping

With the development of the sematic web, more and more applications use ontology. Ontology is a clear conceptualization of specifications, it aiming at achieving a certain degree of knowledge sharing and reuse, and improving interoperability, the system communication and the ability of reliability. But a large number of heterogeneous ontology generated since the characteristics of generating ontology: subjectivity, distributivity and autonomy. This obstructs the semantic interaction between the systems, and violates against the original intention of the ontology, Ontology isomerism became one of the serious obstacles of the ontology’s application. Ontology mapping is the main method to solve the heterogeneous ontology problem, it realizes the mapping by judging the semantic relationships between entities of the ontology.Based on the analysis of the existing ontology mapping methods, this paper puts forward a kind of efficient, automatic multi-strategy method for ontology mapping, and updates each strategy respectively. The main content as follows:(a)Put forward a kind of method based on simplifying the ontology mapping concept set, which is efficiency and multi-strategy. Analyze the ontology in its language, individual, structure characteristics before the mapping, and then use each mapping strategy in proper order. And use the AHP and the characteristics of the ontologies to get the weight of each result automatic, improve the quality of mapping by combining the results based on every strategy.(b)Summarize the existing ontology mapping methods into three kinds:based on language, based on individuals, based on structure, for the language strategy, consider the weight changes following the depth and count of child node of the concept node in the tree of WordNet; and design the strategy based on individual and structure.(c)Realized the prototype system according to the method in this paper, did some test with the OAEI’s standard test-set and compared the result with the existing mature ontology mapping systems’and the traditional multi-strategy methods. The result shows the system has a good performance at precision and time-using.