A session identification method of Web user based on K-means algorithm
Xiao Ping
COMPUTER MODELLING & NEW TECHNOLOGIES 2013 17(5B) 14-18
Department of Network Crime Investigation, Criminal Police College, China
The session identification is an important work in the early stage of analysis of behaviour, which has a decisive impact to find out the behaviour characteristics. After analyzing these common session identification methods, we put forward a kind of optimization method to identify Web user session based on K-means algorithm. We compared the method proposed by this paper with other two methods including θ equals thirty minutes and the session identification based on time distance in three aspects: the number of session, the value of absolute evaluation function A(h) and the value of relative evaluation function R(h). It shows that the session identification method proposed by this paper can identify the real user session more completely.