Intelligent Question Answering System Based On User Design And Research

With the development of internet technology, there emerge a lot of Question Answering Systems on network. However, most of those systems only concern knowledge answering, which consider little about the importance of people as the subject of this activity and also ignore their demand for knowledge management.Based on the Hierarchy of Needs Theory and Collaborative Learning Theory, this thesis has built an intelligent answering system taking the computer science as the whole knowledge and involved the characteristics of social network. This system conducts social activities of knowledge answering through social network. According to the features of social network, the system could effectively manage the content and structure of knowledge of users. In this case, an intelligent system whose centre is users and whose activity content is knowledge answering has been formed.This system divides the whole knowledge of computer science into multiple parts and categories questions according to those parts above. Users are the subject of this system, and those partial knowledge and concrete questions are the attributes of subject of this system. The system mainly includes the direct and indirect interactive activity of subject&attribute. The direct interactive activity of subject&users is consists of users’questions, users’ searching for answer, users’choosing of knowledge and questions interesting them. All of those activities above would be recorded at users’data information, building users’ knowledge content and structure in this system. And the indirect interactive activity is the extension of direct interactive activity. For instance, users could expand their knowledge scope through one part of knowledge recommended by another part; users could understand their own questions deeply through related questions recommended by certain ones.The system is based on B/S architecture. It includes three modules:the sentence processing module, the retrieval and matching of information module, the recommendation and attention of the information module. The sentence processing module puts forward to the maximum matching word segmentation algorithm; the retrieval and matching of information module retrieves by ways of full-text method firstly, and then filters results as well as returns the most accurate results; the recommendation and attention of information module uses the keyword weight algorithm to conduct orders.