Research On The Technology Of Personal Information Retrievl In Search Engine

With the rapid growth of web information, search engines have become the main tools of information retrieval. The existing majority of search engine have the shortcomings of providing the same results to the different user's retrieval requisition, failing to reflect the user's true interest. Research on personalized information retrieval has become an important area.On the basis of studying the personalized search engine and related web mining technology,the TFIDF algorithm based on the word frequency is analyzed deeply, which neglects the relevance between the documents and the user interest., then puts forward a concept of page relevant weight with the web mining technology and the thought of relevant feedback by studying the correlation between the web pages and the user interests. A kind of improved weighting TFIDF algorithm is proposed based on the page relevant weight and the TFIDF algorithm. By analyzing the documents structure, the user's behavior information as well as the user's evaluation to the documents, designs a kind of three-layer strategy for users interests mining based on the weighting TFIDF algorithm to construct and update the user interest model real-time. Analyzing the flaws of the popular relevance ranking algorithm based on cosine of the vectors angle, proposes an improved relevance ranking algorithm based on the user interest model, with which to filter and rank the documents according the user's interests and hobbies in order to improve the precision rate. Designs and realizes a personal information retrieval system. In this system, the inquiry expansion is realized by combining the user's interest model and the inquiry keyword the user has inputting, and the documents for personalized filtering and ranking is finished by matching the searching results and the user interest model. Under the condition of guarantee recall rate, improves the system's precision rate, and realizes the goal of personalized information retrieval. The effectiveness of the algorithm was proved by the results of the experiment.