Syntactic structure feature analysis and classification method research

Characteristics and difficulties of Chinese natural language processing, as well as the appropriate language to deal with the basic resources relative lack of some mature foreign technology and research results can not be applied directly to the Chinese question answering system. To this end, the question answering system for Chinese facts, presents a new extraction algorithm based on the characteristics analysis of syntactic structure and classification of the answer, the answer extraction problem as a classification of the candidate answers, upcoming candidate answer classified as correct and error categories. First, the method according to the types of problems corresponding to the type of candidate answers extracted from the text fragments candidate answer in the sentence simple features and syntactic structure characteristics; then take advantage of these features of the training classifier; trained classifier to distinguish the candidate answer for the correct answer.