Thai Sentence Similarity Calculation Research

Thai sentence similarity calculation is one of the problems in the field of information processing.With the implementation of the national "one side" strategy,China's economic exchanges with Southeast Asian countries are becoming more and more close,and Thailand is one of the important countries in Southeast Asia,Natural language processing research is very important.Thai sentence similarity calculation in a number of areas have a wide range of applications,especially the Thai machine translation and expert system,similarity calculation is an indispensable part.During the research of Thai sentence similarity computation,the paper studied a Calculating the similarity of sentences from 3 features:Keywords semantics,syntactic information and syntactic of WordNet corpus.According to the Strengths and Weaknesses of calculate sentence similarity,according to three features,The calculation method of the similarity of sentence sentences has been improved in the paper,and we have proposed a better way to calculate the properties of three sentences.And finally,we use the application.of Thai sentence similarity Calculating system.The main of Calculating methods in this paper are as follows:1.Study and analyze similar sentences based on the vector space model and study the Thai word wrap method.Then used to create a test system to compare Thai sentences.2.analyze in The semantic similarity of Thai vocabulary and its calculation method.The sentence is composed of vocabulary combined with specific grammatical rules.Therefore,the similarity calculation technique of words is the basis for the study of sentence similarity calculation.Therefore,it is very important to study the similarity calculation of words..3.Study and analyze the technology of calculating the similarities of existing texts,focusing on the use of WordNet calculating method as a meaningful and analysis meaningful of Thai language.4.We design and implement a Thai sentence similarity computation system and do some experiments in this system to test the infallibility of sentence similarity calculation method proposed by this paper.