Ontology Mapping Based On Structure Quantification

With the application prospect of ontology in fields like knowledge management,semantic Web and artificial intelligence becoming clear,more and more ontology have been developed.At the same time,the heterogeneity problem between different ontologies resulting from different researching or application purposes has received more and more attention.As ontology heterogeneity obstacles knowledge sharing,reuse and interoperability,ontology mapping which is a fundamental way to solve the problem of ontology heterogeneity has become the focus in the field of ontology research.Currently,the Ontology mapping research field has formed a methodology system covering syntax based method,structure based method,example based method and external resources based method as well as multi-strategy ontology mapping methods.However,the existing ontology mapping methods have the common problems that the methods' accuracy is not good enough and they do not take the most of the potential semantic information.This dissertation firstly defines and quantify three structural information of ontology concepts,namely depth,width and density,and then research the methods that converse the quantified structural information to the semantic relevance between the father and the son concepts.Then propose the method to use the semantic relevance to get ontology mapping by build Common Character Vector.By combining the Common Character Vector and the semantic relevance of a concept pair,the calculation of similarity of concept pairs can be conversed to calculation of similarity between two numerical vectors so that the result of ontology mapping can be achieved.This ontology mapping method solves the problems that how to extract structural information of an ontology and how to use it to get ontology mapping.Experiment results show that this method show a better result of discovering semantic relationship between two different ontologies.Especially with the help of structural information,the method can discover potential semantic relations of the concept pairs which contain two concepts different in form but consistent in meaning or different in meaning but consistent in form,so that the precision can be improved.It is proved by the test data that the approach got approving precision,recall and F-measure.And an ontology mapping system based on the proposed method of this study is built to show the practicability.In conclusion,based on the structure of the ontology,this study proposes a new approach of ontology mapping to solve the heterogeneity problem between different ontologies,and innovatively proposes a method to quantify ontology structure information and converse the calculation of similarity of one pair of concepts to the calculation of similarity between two vectors of two numerical vectors.This study can enrich and improve the method system of ontology mapping.