The Research Of Understanding Question Sentence And Personalized Recommendation Algorithm In The Automatic-answering System

In recent years, research in the field of natural language processing is generally being concerned by lots of people. Automatic anwering system is becoming a hotspot of the field. With the rapid development of Internet technology, more and more people tend to gain knowledge and find solutions from the Internet. But with the improvement of users' requirements, the traditional automatic answering system cannot easily meet people's needs. On one hand, this is because the machine's ability to understand is limited, which cannot understand questions in the form of natural language. On the other hand, users hope the system can provide personalized recommendations according to their different interests.With the expansion of school's enrollment, Teachers cannot answer everyone's questions. So the field of automatic answering system has become a new hotspot.This dissertation focuses on the problem of personalized recommendations and how to understand question sentences, which are the two problems in the traditional automatic answering system, and did a lot of research on it. The main works and Innovative points are as follows:(1) We build the automatic answering system as a platform in which we used the information extraction technology from the the search engine, so in this way, the system can update its knowledge base automaticly.(2) We improve the understand module of the system by raising a concept called "the hierarchical sentence template" based on the research on a lot of question sentences. And the hierarchical sentence template improved the accuracy of understanding question sentences.(3) We used a new way to design the users' interest model which is called "the additional weight method". We combine two personalized recommendation methods to recommend in the automatic answering system which is used in the field of education. The system has the ability to recommend to the students what they will be interested in and the people who always ask the same questions with.