E-Commerce Website Structure Optimization Based On Web Usage Mining

With the development of web technology, e-commerce website developed rapidly. Nowadays, almost every enterprise has its own website. Since possessing a well-structured website has been the base of the enterprise operates E-commerce successfully, also been an important factor to promote the enterprise's further development. Then how to improve the website structure's rationality has been one of the crucial problems which the website operator must solve.The above problem can be solved by the website structure optimization technology, which optimizes the website structure through using data mining technology to find the existing structure's deficiency and amend it. In the various existing website structure optimization technology, based on the web usage mining technology is a significant research direction. User is the only estimator of the website structure's rationality. His accessing behaviors are recorded into the web access logs on web server. These logs include some information, such as user's access URL, access time and user's IP, etc. Web usage mining technology can take full advantage of these information to find the user's access patterns, and guide the e-commerce website structure optimization work by analyzing and researching the rules behind the web logs.Through the research of web usage mining technology and website structure optimization technology, an e-commerce website structure optimization system (called: WSOS-WUMbased) was designed and completed in this paper. Using this system, e-commerce website's administrator can make use of the users'access logs to find the existing structure's deficiency and adjust it, made the website structure optimization. The WSOS-WUMbased was composed of two functional modules: information processing module and structural adjustment module.The main task of the information processing module was: firstly, collect the data as much as possible for the system, and transform these web data into a suitable form. Then the appropriate algorithm was chosen to deal with these data. In this system, hidden Markov model was used to model the user's access behaviors and mine the user's access pattern.The main task of the structural adjustment module was: firstly, the existing e-commerce website structure's deficiency was found by using the user's access pattern. Then a suitable method was chosen to optimize the e-commerce website structure. In this system, the method of adding new hyperlinks was used, in order to ensure the original website performance at steady state.