Research of Web User Access Pattern Mining Based on Session Classification

WU Hai-yan,ZHU Jing-jun,GAO Guo-zhu,CHENG Zhi-rui
2010-01-01
Abstract:With the rapid development of Internet,web users are increasingly diversified.The access patterns of different kinds of users are different.This paper presents a method of mining web user access pattern based on session classification.This method divides all the user sessions into three categories,which are human user sessions,crawler sessions and resource-download user sessions.Then mine the access patterns of each kind of users respectively.Session classification can improve the efficiency and accuracy of mining.This paper focuses on mining the access pattern of human users.It presents a method of transaction identification based on users′ access-path-tree,and makes some improvements on PrefixSpan algorithm.This mining method achieves good results in the experiments.
What problem does this paper attempt to address?