Analysis of Fuzzy Clustering Techniques Used for Web Personalization

Web personalization aims to provide content and services tailor-made to the needs of individual users usually from the knowledge gained through their (previous) interactions with the site. Typically, an access behavior model of users is learnt from the usage of the Web site which is then used to provide personalized recommendations to the current user(s). In this paper, we present a detailed qualitative as well as experimental analysis of various fuzzy clustering techniques used for mining usage profiles. We discuss their algorithmic strategies, requirement of input parameters, noise handling capacity, scalability to large datasets and similarity of partitions. We validate our claims through experiments using a large real life dataset