Research And Implementation Of Chinese Ontology Mapping Based On Integrated Similarity Calculating

Ontology is the important foundation of reflecting semantic of information resources and its application requirements are wide. Nowadays, the research in Chinese ontology mapping is limited. The similarity calculating accuracy of ontology mapping is a little low. And the problem of ontology heterogeneity has come out since there are not unified standards for ontology. This paper focuses on the research in Chinese ontology mapping and it explores the effective methods for solving ontology heterogeneity. The research is of great significance and practical value, and the main points are shown as below.1. Proposing a Chinese ontology mapping framework based on integrated similarity calculating. The framework includes ontology preprocessing, integrated similarity calculating and ontology selecting. The integrated similarity calculating includes conceptual similarity calculating and structural similarity calculating. The combination of two calculating methods can improve the accuracy of ontology mapping, which is very low in single similarity calculating.2. For the conceptual similarity, this paper proposes a similarity algorithm based on primitive hierarchy. The main idea of this algorithm is:firstly, it needs to calculate primitive similarities, secondly, it needs to weight and sum all the primitive similarity values according to the importance of primitives, and then get the conceptual similarity at last. Because primitive similarity is the basis of conceptual similarity, it can increase the accuracy of conceptual similarity by means of improving the algorithm of primitive similarity.3. For the structural similarity, a pair of elements would be similar in structure if the conceptual similarities of their neighbors are high, which are for two ontologies to say. According to this, the paper proposes a similarity algorithm based on neighboring relationship, and the algorithm has improved the structural similarity calculating. 4. This paper implements a Chinese ontology mapping system named COMBIS (Chinese Ontology Mapping Based on Integrated Similarity), which is based on integrated similarity calculating. The paper designs an instance of ontology mapping and utilizes it to COMBIS system. The methods are proved of effectiveness via analyzing the experimental results.