Prediction of user’s type and navigation pattern using clustering and classification algorithms

Abstract

World Wide Web (WWW) means huge amount of web pages and links that provides massive information for internet users. The growth in websites has become more complexity and size of web contents is more abundance. A web usage mining techniques is used in web server log for extracting a user behavior. Three types of user behavior include frequent user, synthetic user and potential user. The objective of this paper is to predict the potential user and it’s navigation pattern from the web log files. The process of web mining works in three main phases such as data pre-processing, classification of users and pattern discovery. This paper deals with clustering techniques for pattern discovery such as path prediction, page gathering, fuzzy clustering, ant-based clustering and graph partitioning, etc.. Comparative study of these techniques gives best results for predicting future visit of potential user in web server log. Among all clustering algorithm, fuzzy clustering algorithm gives 98% accuracy for predicting potential user navigation pattern which is higher than other techniques.