Analysis of Web Usage Mining Using Various Fuzzy Techniques and Cluster Validity Index

The internet is a large source of information and there is a vast growth in the degree of information each moment. The large number of web users are also increasing every day. To decrease web users browsing time, a lot of studies are going on. In Web usage mining technique mining strategies are performed in a proxy server dataset to detect the behavior of web users. Clustering plays a key role in a large scale of applications like web log dataset investigation, CRM, advertising and marketing, scientific diagnostics, computational biology, and lots of others. Clustering is the set of related data items. The key issue for clustering is a few sorts of measures that can decide whether two objects are relative or distinctive. This paper describes various validity measures including partition coefficient and partition entropy. Experiments perform for the various fuzzy clustering techniques such as Fuzzy C-Means, Fuzzy Possibilistic C-Means and Modified Fuzzy Possibilistic C-Means Clustering. These methods are developed and examined for proxy log dataset. Finally algorithms experimental outcomes are analyzed. Experiential result precisely denote that the clusters formed using proposed MFPCM method is much better in terms of various validity parameter compared with k-Means, Fuzzy C-Means, Fuzzy Possibilistic C-Means and Modified Fuzzy Possibilistic C-Means Clustering algorithms.