Concept-based clustering of clickstream data
Anurag Narayan Banerjee · 2000
Determining the type of a user based on his interaction with a website is a key problem in web usage mining. Absence of proper cookie information makes the task more difficult since one then has to extract exact session information for the anonymous users before performing any advanced analytics such as user characterization. In this paper, we first discuss two ways of session extraction from weblog data. Then we propose a novel and effective algorithm for clustering webusers that takes into account both the trajectory taken through a website and the time spent at each page. Results are presented on weblogs of www.sulekha.com to illustrate the techniques.