Abstract: The process of web usage mining is implemented in an attempt to discover the patterns of user browsing from the web log data available on the server side of almost each and every website or in the user's cache. By analyzing the browsing behavior of various users, next web page predictions can be made which is an important aspect of most of the websites around these days. But the prediction of future requests comes with its own set of issues which make it unreliable at times; there are some concerns with accuracy and efficiency. In this paper, we have designed a custom algorithm for the Clustering process. The main aim of this algorithm is to provide more efficient and accurate results as compared to the other Clustering algorithms.

Keywords: Web Usage Mining, Recommendation, Prediction, Browsing Patterns, Patterns, Clustering, Preprocessing.