Conceptual Semantic Similarity Calculation Based On WordNet And Its Application Research

With the development of the Information Technology and Artificial Intelligence,more and more people get information from Internet.So,it is necessary to improve the accuracy of retrieval information.As one of the key technologies of natural language processing,Similarity calculation can filter the information efficiently,in order to improve the quality of information retrieval,we can improve the accuracy of similarity calculation.Similarity calculation includes the calculation of semantic similarity and the calculation of sentence similarity.The improvement of the performance of similarity algorithms has a crucial impact on the development of its application.In this paper,concept semantic similarity computation and the sentence similarity calculation are studied in detail.The MICS model of the concept semantic similarity is proposed and is applied to the calculation of the sentence similarity.The test results show that the MICS model has better performance and improves the accuracy of similarity calculation.The research work is as follows:1.Expounds the background and the significance of the similarity method,and discusses the necessity of this research.Analyzing the research progress of concept semantic similarity algorithm and sentence similarity algorithm.2.Introduces the knowledge of the WordNet semantic dictionary.Focuses on the development,the content structure and the semantic relations of the WordNet.Introduces the WordNet version and its structure used in this paper.3.According to the advantages and the disadvantages of the usual algorithms of semantic similarity,the MICS model is proposed.The model is based on the IC(information content)model,using the conditional probability between neighboring concepts to weight the edges,and mutual information is used to express the semantic similarity of two concepts.The algorithm takes into account the density,the depth and the path of the concept in the hierarchical tree,and combines the algorithm based on the IC and the algorithm based on the path.Then the MICS model is tested and analyzed.The results show that the proposed algorithm has better performance.4.Based on the detailed analysis of sentence similarity calculation steps and some algorithms,the MICS model is applied to the calculation of the sentence similarity,and the MICS model is proved that it has a good performance.5.Analyzing the inadequacies of this paper and the problems that need to be solved.Finally,the future research work is prospected.