The Research of Chinese Semantic Similarity Calculation Introduced Punctuations

So far, most Chinese natural language processing neglects the punctuations or oversimplifies their functions. To improve the efficiency of Chinese similarity computing, this paper gives a Chinese similarity computing system model in accordance with the problems of Chinese sentence similarity computation aspect. This model is a combination of punctuations and traditional similarity computing. Comparing with Cosine-based Similarity calculation, the sentence similarity calculation based on word shape and word order, and Sentence similarity calculation based on semantic, this model makes up their insufficiency to a certain extent. The experiment shows that this model has a higher rate of accuracy in Chinese similarity computing.