Research On The Intelligent Question Answering System Which Calculation Is Based On Chinese Question Semantic Relevance In The Restricted Domain

With the rapid development of computer technology and artificial intelligencetechnology, Intelligent question answering system more and more appear in People’sDaily lives and Play the role of more and more important. Users in the pursuit ofanswers’ accuracy in the same time, more and more concerned about the authenticityof intelligent question answering system.So intelligent question answering systemshould have answer returned high accuracy’ feature and strong authenticity ofhuman-computer interaction’ feature. Therefore, this paper based on the results ofprevious studies, Focus on the right research on the intelligent question answeringsystem which calculation is based on chinese question semantic relevance in therestricted domain，Achieved a algorithm which calculate based on the questionsemantic representation and the domain keyword tree.The main contents of this paperare:(1) Research on domain keyword extraction technology,compare with SeveralDomain keyword extraction technology, Propose a Domain keyword extractiontechnology which suitable this paper, To achieve the establishment of areas ofkeyword dictionary.(2)Research on currently available Semantic relevancy algorithm and questionssimilarity algorithm,Achieve Semantic relevancy calculation methods which moresuitable this paper,Achieve Semantic relevancy calculation For statement.(3) Achieve research on the intelligent question answering system whichcalculation is based on chinese question semantic relevance in the restricted domain,Verify the correctness of the theoretical methods.The innovation of this paper are:(1) Improve Domain keyword extraction technology.In this paper, the termextraction of mining areas TF-IDF word filtering technology combined with activeareas of technology for extracting word, This method not only takes advantage of thecharacteristics of TF-IDF Technology,but alsospecially added active word filter,make the domain keyword extraction more accuracy.(2)Proposed a new method of semantic relevancy computation.In this paper,Semantic relevance calculate based on the question semantic representation and thedomain keyword tree, Mainly uses existing algorithms and the improvement of itscorresponding,and calculate the semantic relevancy combine the domain keyword treewhich establish in this paper. Not only improve the accuracy of the intelligentquestion answering system, But also to make the system more in line with the contextof intelligent quiz questions and answers.Above the results,in brief,this paper designs and completes the prototype ofchinese intelligent question answering system based on related information ofshandong finance and economics university.experimental proof that question semanticrelativity calculation algorithm question semantic representation and keyword treebased on accuracy and efficiency are presented in this paper has an ideal effect.