An Improved Concept of Semantic Similarity Computation Algorithm based on Domain Ontology

Abstract

To effectively solve the problem of semantic similarity between concepts, the existing concept semantic similarity computation methods were studied and an improved concept semantic similarity computation algorithm based on domain ontology was put forward. In the process of computing concept semantic similarity the algorithm not only considered the basic relationships but also the custom relationships between concepts. In addition, the algorithm also took into account the concept of properties, and the concept of instance impact on semantic similarity computation. All these measures make the effectiveness of the algorithm in computing concept semantic similarity improved. The example showed that in the aspect of concept semantic similarity computation the proposed algorithm is more effective than the existing algorithms. Introduction To the problem of concept semantic similarity computation the domestic and foreign scholars have carried on the related research and achieved some results. Richardson determined the weights of the concept according to the node density, depth and strength and then commutated the semantic similarity between concepts based on the weights [1]. Yuhua Li used concept depth, density and length between concepts to construct a nonlinear function, and then used this function to commutate semantic similarity between concepts[2].Lin-tao Lv presented a computing concept similar model based on context[3].Literature[4] proposed a concept semantic similarity computation method based on the number of the concept properties. According the upper and lower relationship between concepts and other relationship Jie Chen gave a new computation method[5].Mei-rong Yang proposed a concept of semantic similarity computing model based on the main similarity between concepts[6] etc. The above semantic similarity computation algorithms only considered certain aspects of the impact semantic similarity. That make the algorithms to achieve good results in certain applications. So an improved semantic similarity computation algorithm was proposed. The algorithm took into account the basic relationships between concepts, custom relationships between concepts, the properties of the concept, the instances of the concept and other factors influencing in the process of computing the semantic similarity between concepts. That effectively improved the algorithm's effectiveness and versatility.