WEB PREDICTION USING FCM AND KFCM ALGORITHM

Abstract

In present days, Internet is playing such a significant role in our day-to- day life. We have witnessed the evermore- interesting and upcoming publishing medium
is the World Wide Web (WWW). The rapid growth in the volume of information available over the WWW and number of its’ potential users’ has leads to difficulties in providing effective search service for users’, resulting in decrease in the web performance. Web Usage Mining is an area, where the navigational access behaviour of users’ over the web is tracked and analyzed. So that websites owner can easily identify the access patterns of its users’. By collecting and analyzing this behavior of user activities, websites owner can enhance the quality and performance of services to catch the attention of existing as well as new customers. Web prediction is a clustering problem which attempts to predict the most likely web pages that a user may visit depending on the information of the previously visited web pages. In this research work focus is on web page prediction. For web page prediction we use algorithms FCM and KFCM and comparison between them on the basis of conditional variance parameter. But to remove the fuzziness and to reduce the time complexity we have introduced a term
Gaussian under KFCM and named it as GKFCM.