Method Of Computing Ontology Semantic Similarity Based On Weighted And Its Application

Recent years, with the rapid development of computer science and technology, ontologystudy has become a hotspot, ontology has a very wide range of applications in many areas suchas the Semantic Web, E-commerce, information retrieval, knowledge engineering, computerinformation, knowledge processing platform and artificial intelligence. However, with theincreased number and size of ontology and the diversity and autonomy in ontology building andusing，many heterogeneous problems produced, which greatly affect the efficiency of ontology,it’s the most important issue to be resolved currently. Ontology similarity calculation is the keyto solve the problems of heterogeneous among different ontologies, it’s an important foundationwork for information exchange and processing among ontologies, how to integrate and makefull use of existing resources need ontology semantic similarity algorithms.The main contents of this paper is divided into two aspects: First, propose a method ofontology similarity calculation based on weighted, and verify the effectiveness of the algorithm.Second, ontology similarity calculation based on weighted is applied to three-dimensionalmodel area. The retrieval method is more effective than traditional method in three-dimensionalmodel retrieval, How to integrate and make full use of the existing ontology resources is neededfor ontology similarity computation support.Detailed work steps are as follows:(1) Build functional ontology according to the function of three-dimensional model.First, extract the function information of three-dimensional model，build the prototype offunctional ontology library with the help of experts in three-dimensional model area, usingWordNet compute similarity of two words to expand three-dimensional model. Eventually，build a ontology library of three-dimensional model.(2) Propose a method of ontology similarity calculation based on weighted. In the searchof ontologies, no longer just use tools such as WordNet、Hownet directly get the similaritybetween the two ontologies, the concepts and properties of ontologies are taken into account,set the corresponding weights according to the depth and width of ontology structure， Finally，the top-down method of ontology matching is used to complete the calculation of similaritybetween two ontologies.(3) A method of ontology similarity calculation based on weighted is applied tothree-dimensional model retrieval. On the basis of the functional ontology has been completed, compare the target ontology and resource ontolo gy in ontology library by the method ofontology similarity calculation based on weighted, at last，the relevant experts set threshold tocomplete the three-dimensional model retrieval.(4) Developed a simple simulation system to achieve three-dimensional model retrieval.Designed user interface, allowed users to operate simply, users can select different thresholds orweights，the system achieved ontology similarity calculation in three-dimensional model,applied it in three-dimensional model retrieval, the experimental results depending on thecircumstances is effective.