Personalized Recommendation Technology In Agent Answering System-based Applications

The Question Answering System (QAS) plays an important part in the web-based teaching. In recent years, the research and application of the QAS is one of the important topics noticed and discussed. But there are still some promblems in most of the QASs, such as the simplify mode of question answering, the fussy manipulation and the restrained application, even the large numbers of resource not be fully used and the severe redundance of the data,which lead the systems to be lack of applicability and intelligence, and difficult to satisfy the requests of the personalized tutoring and inspire the enthusiasm of the learners. As a result, the efficiency of prediction of the E—Learning is not achieved.The paper introdues the agent technology, analyses and designs the system based on agent technology. Through the research for the application of the collaborative filtering technology in the QAS, the system gets intelligence and more applicability. On the other hand, it can service users differently according to their level.In this paper, the technology based on agent and the conception of personalized service are introduced. In addition, the clustering analysing method about the user in Data Mining and the collaborative filtering technology are also discussed. A model of the QAS is presented, of which the main function including User Agent and Working Agent which is composed by Searching Agent and Personalized Agent is analysed and designed in detail. The function structure and the work flow of the modle are also given. The key technology about how to implement the Personalized Agent is the K—Medoids arithmetic which is researched and improved to make the recommendation more fitter to the user's needs.