Research On Personalized Search Engine Based On Web Data Mining And Information Classification

As the Internet is developing drast rapidly, on-line data shows an explosive growth. The early Internet search engines just accumulated the data by the way of indexing, and displayed the same data to different user groups. This may lead to the result that general search engines are unable to meet people's specific search needs.In this theis, firstly, the working principles and architectures of search engines are introduced. Search engine related technology is described. Information performance evaluation, vector space model and the PankRage algorithm are summarized. Key issues and technology of implementation of the personalized search engine are discussed. Web data mining and features of user behavior characteristics mining are described. Based on the existing search engine technology, theoretical methods to implement personalized search engine are proposed. Secondly, web-based vector space model and Bayes for automatic classification is combined with individual-based mining model:1. It is based on the behavior of individual mining and the realization of personalized content mining; 2. Users directly enter their own key words of interest to search the relevant information in a huge database of search engines. Thirdly, implementation of the information classification search engine based on Nutch is described. By using the new search engine can give different search results according to different user's personal requirements. Finally, personalized search engine technology and research directions are summarized.