Research On Mining User Access Mode Based On Web Logs

With the rapid development of the network and information technology, the application based on Web has covered all aspects of social life, so the data on the Web is usually huge. In these data, compared with the web structure and content, the using mode of users is more remarkable. By obtaining the access mode of users, we can realize a series of application from optimizing the design of websites to improving the customer relationship. Designing and modifying the structure and layout of websites according to the behavior patterns of users can let users visit interested pages in the shortest time and improve service performance. Understanding and analyzing the access behavior of users can find potential customers and keep users presence. By mastering the access behavior of users, group decision makers can design the goods catalogue more purposefully and improve the accuracy of business decisions. Finding the access pattern of the individual user can recognize the user's interests, hobbies, habits and needs, establish the personalized user model, provide the content and service more personally.The information of usage modes is usually revealed in Web server logs. Web server logs records the mutual reacting information between users and servers which reflects all movement by users. Mining and analyzing Web logs can obtain access patterns and useful information including hobbies and interests so as to understand the access behavior of users.Basing on the method and the process of Web usage mining and using Web server logs as data sources, the paper aims to mine frequent access paths of the individual user and group users, in order to find out the access modes of web users.As for mining frequent access paths of the individual user, the paper introduces two representative association rule mining algorithm in detail, improves the algorithm by adding interesting measurement factor from the aspect of enhancing usefulness of association rules. As for group users , the paper analyzes the process of user clustering in detail, and uses UBPC to mine access patterns of users.Finally, the paper proposes a model of user access patterns mining system, introduces the function of each module, does the experimental analysis by combining the concrete data to illustrate.