The Research And Implement Of Algorithm On Web Usage Mining

With the rapid development of network technology, information online increases on an exponential basis. Web becomes the most important carrier that stores, releases and obtains information information. Web mining is the application of traditional data mining technology in the web environment, finding out the user's browsing pattern and relevant web pages. The purpose of study on the mining is to find out the action pattern of browsing online, improve the websites structure and hyperlink structure and the service quality of websites, provide support for the policy decisions on the clients' relationship management in e-commerce. The thesis focuses on the search of mining algorithm for web including the following contents:Firstly, the theoretical knowledge on data mining and web mining is introduced, in addition to the analysis of research focus.Secondly, the algorithm for the maximum forward session path is designed on the basis of websites access preprocessing. Then, a new session construction method, an elicitation method based on time and citation, is introduced, which not only takes advantage of features on time, but takes into account the user's browsing habits. It, to some extent, brings users convenience on the mining for the frequent access pattern.I put forward a FAP-Mining method, which can be used to find out the user's access pattern on different types of websites, and obtain the frequent access pattern according to the threshold set by experts.At last, an experimental system for web application mining is developed, which implements the session construction method based on time and citation. FAP-Mining algorithm is implemented in access pattern module.