Study And Implementation On Personalized Recommendation Techniques Based On Proxy

With the coming of information age, network technology develops rapidly. As a result, many research problems about E-service technology appear one after the other. Among them, the Web personalized recommendation technology has become a key research issue, which can provide personalized services for Internet client expediently.Web personalization means that a Web site can customize browsing content or give browsing suggestion for users dynamically according to their browsing favors.This Thesis focuses on how to design and implement a Web personalized recommendation system based on proxy, which name is Smart Web.According the analysis and comparison about the architectures of different kind recommendation systems, the personalized recommendation system based on Proxy has advantages which make it can be put into practice without increasing the burden of client and Web station. This Thesis introduces how to use Web usage mining technology to implement this recommendation system in detail.The Smart Web system is composed of two parts: the off-line mining sub-system and the on-line recommendation sub-system. In order to have the system make good progress in its capability, the recommendation system makes use of the web site structure and clustering analysis methods to filter association rules discovered in off-line mining sub-system, and then constructs an index d ata s tructure according the filtered rules to enhance the speed of recommendation in the on-line recommendation.In the last chapter, the thesis evaluates the capability of SmartWeb system in speed and efficiency. The experiment results show that the SmartWeb system has better precision of recommendation and faster speed; therefore, it is feasible to put this recommendation system into practice.