A Method For Calculating The Similarity Of Chinese Sentences

In the Chinese information processing, Chinese sentence similarity computation used for document summary, voice systems, data mining, text fields etc. It is a very critical issue,and it has long been research hot and difficult. In web search, semantic correlation between sentences and web pages refers to web pages that can help users to meet the needs of users. The traditional correlation calculation is mainly based on the keyword matching, which is difficult to determine because of the key words, and it often result in the calculation of the Chinese sentence similarity is not accurate.This paper in the research of Chinese sentence semantic similarity calculation process is considered a sentence is an expression of complete information,and each sentence in the composition of existing semantic dependency relations. On this basis, establish a semantic dependency tree. Through the Chinese sentence semantic dependency tree analysis, we discovered the existence of an exponential relationship between the tree layers Chinese sentence semantic dependencies. On the basis of this relationship, the paper gives the calculation method and calculation formula. The establishment of Chinese semantic dependency relation tree accuracy is not high, but also a combination of similarity calculation method of frequency probability.The new idea of a semantic relativity calculation. In traditional text similarity technology go into the semantic dependency tree, characteristics term weighting, information entropy technology. In this paper, we use the proposed algorithm to calculate the similarity of semantic correlation dimension data set of Chinese sentences, and verify the accuracy and effectiveness of the proposed algorithm. Experimental results show that this algorithm is more accurate than the traditional algorithm.