The Analysis Of User Interest Webpage Based On Quantitative Calculation Of Browsing Behavior

Web has become an important way to get information by people, with the increasingly growth of Web information, people have to spend a lot of time to search and browse the information they need. Search engines have become the most widely used information retrieval tools. However, nowadays, the most of services provided by the search engines can not completely make the users satisfied, especially for a particular user's personalized service. How to accurately collect user interesting web pages is the important basis for the varieties of personalized services technology or system research and development, meanwhile, it is the key of personalized service which determines the quality of the personalized service provided by system. This paper is based on the recommendation system realization of a complete personalized search engine services, and it completes the following research works:①Make the detailed analysis of the importance of varieties of browser behaviors from the user's interest view,. The behavior of user browsing web pages can reflect the user's browsing interesting at a certain degree, using these on the field of personalized service, we can find that there is a certain relationship between the user's browsing behavior and the user's interest degree of web pages. This paper analyzes the browsing behavior that impacts user's interests , and proposes the method of considering user average behavior in the behavior analysis according to the disadvantages of present user behavior analysis.②Propose the quantitative calculating method of web pages interests of different browsing behavior, and design a kind of interest website extraction technology which is adaptive for parameters. It also propose the interest website extraction method which is mainly automatic extraction and auxiliary manual extraction. Adopting the technology of parameter self-adaptive and non-normal web pages automatically removing, take full advantage of the interest web page sets and non-interest web page sets got in the course of extraction to gradually modify extracted parameters in order to achieve the purpose of correctly capturing user interest web pages, which supplies the reliable, high-quality data for the following web mining.③Thirdly, this paper put forwards a method to determine the border interest page using search terms of user. It proposes a method to capture the search terms for search terms commonly used, and establishes the search word dictionary using the search terms. Moreover, it revises the formula of calculating interest through the method of combining the search terms with the contents brown by user, and has improved the identification accuracy of interest page nearby critical point extracted in the automatic extraction methods.④For the several methods proposed above, this paper dose the experimental analysis based on the personalized search prototype system to verify the effectiveness of the methods. And preliminary experiments show that the rate of exactness and the rate of recall extracted by interest page are high, and has reached the desired objective,which can improve the quality of service of personalized search.Nowadays, personalized service is gradually becoming a hot spot whether in academic research or in commercial applications. In this paper, the extraction models of interest pages and the expansion method of search engine can be used in the field of personalized information services, search engine expansion, customer information management, e-commerce, data mining, and so on .