Chinese Question Understanding Technology Research And Application In The Field Of IT Question Answering System

Question answering system, as the new generation of intelligent search engine, allows users to ask question by means of natural language, and can supply more accurate answers. Comparing to the traditional search engine which is based on the keyword, question answering system can provide an accurate and concise answer to a natural language query. Question understanding is a very important component of question answering system. The accuracy of question understanding directly affects the accuracy of the ultimate answer extraction. In this paper, preliminary to establish a question’ classification mechanism based question of intent based on How Net; then, according to the semantic classification block feature extract to obtain semantic understanding model; secondly, we proposed multi-level and multi-feature combination sentence similarity calculation method based on sentences and question’s focus identification aimed at the purpose of Chinese sentence; Finally, according to the above-mentioned theory, the establishment of a practical application of the model. The main works in this paper are as follows:Firstly, through analyzing characteristics of Chinese question, we proposed the classification method based question intention. by extracting three aspects of sentence elements including focus of question, syntactic and semantic roles, we divided the individual components of the classification question into event semantic block, asked point block, the other semantic block. and with extracting semantic of each block we establish semantic understanding model of question.Secondly, question extraction method will differ depending upon the different types of question, namely the intention of the study is to extract interrogative. Research background is a natural language question answering system, through a large corpus analysis, and the number of different frequencies in question, we especially divide it into the interrogative type question, negative question, question of with sentence particles. we proposed the targeted extractive mothod about question intentions on the basis of the interrogative type.Thirdly, by analyzing and researching the limitations of the existing question similarity calculation method, the Chinese sentence similarity computation based multi-level feature and fusion is proposed. The main idea is as follows: First, for the interrogative user made, using natural language processing techniques to analyze the problem, the user’s question can be divided into question intention and keyword sets. The second step, we consider the target layer, structural layer and semantic layers, including intent, keyword, sentence length and other characteristics. The third step is to determine the value of the sentence similarity, namely the use of a simple and effective means to obtain comprehensive integration features, and then determine the value of sentence similarity based on integrated features.Finally, in this paper, we put the question technology studied and a new method for computing sentence similarity differ from the existing understanding of technology used in IT field of natural language question answering system. Evaluation results of the automatic question answering system for the IT field shows that the method is feasible and high practical value.