A RSS User Interest Based Personalized Search System

With the development of the Internet, people generally retrieve information on World Wide Web. While, the information is poorly organized and its amount is very huge so there is an urgent problem that people cannot retrieve useful information effectively. More and more people realize that it is faster and more convenient to retrieve information by search engine; however, current search engines are not strong. Professional search that includes personalized content and other functions has increased, of which the update cycle is still on days or even weeks, so nowadays professional search cannot significantly improve update speed.This dissertation discusses a kind of search engine based on user's interest. The search engine captures user's behavior, identifies user's interested information on a variety of learning methods, and then updates user's interest model to fit to user's real interest, which achieves effectively personalized search.Extended as a kind of XML formats, RSS has some characteristics such as syndication and aggregation, which not only allow users to access information with less time cost, but also characterize the user's individual needs. The news and information that RSS users received is timely and formative, so RSS is useful and convenient because of its timeliness and richness of its content.In this dissertation, RSS user's interest is divided into a couple of concepts as"main interest"and"potential interest", and an improved model of RSS user's interest is discussed. Also, a RSS user interest based personalized search system is designed and implemented. As a tool,RSS Plug-in is embedded in the browser. The system is capable of RSS user's apply from user's interest in rapid analysis and classification, and then clusters according to user's interested categories, while RSS user could learn and feedback in time, also, the system updates user's interest model to fit to the changes of user's interest synchronously. So the RSS user interest based personalized search system avoids low precision when user's interest is changing and achieves an efficient personalized information services.