WEB BROWSING FEATURE MINING OF AN ANONYMOUS USER
ZHAO Liang, ZHANG Shou Zhi, and FAN Xiao Feng (Deparment of Computer Science, Fudan University, Shanghai 200433)
Analysing a user's behaviour pattern based on his interacting with a website is a key problem in web usage mining, especially to an anonymous user. First discussed in this paper is how to extract session information from web usage data, and then introduced is the session feature extracting and feature space description. Based on these, a highly efficient web browsing feature mining algorithm of an anonymous user is proposed. This algorithm reduces the computation consumption based on enhancing the accuracy. It may solve these questions well such as change path length, the directivity and dynamic clustering.