Extracting Web User Profiles Using Relational Competitive Fuzzy Clustering

The proliferation of information on the World Wide Web has made the personalization of this information space a necessity. An important component of Web personalization is to mine typical user profiles from the vast amount of historical data stored in access logs. In the absence of any a priori knowledge, unsupervised classification or clustering methods seem to be ideally suited to analyze the semi-structured log data of user accesses. In this paper, we define the notion of a "user session" as being a temporally compact sequence of Web accesses by a user. We also define a new distance measure between two Web sessions that captures the organization of a Web site. The Competitive Agglomeration clustering algorithm which can automatically cluster data into the optimal number of components is extended so that it can work on relational data. The resulting Competitive Agglomeration for Relational Data (CARD) algorithm can deal with complex, non-Euclidean, distance/similarity measures. This algorithm was used to analyze Web server access logs successfully and obtain typical session profiles of users.