Exploring User Navigation with the Synergy of Modified Ant Based Clustering and LCS Classification

Abstract

World Wide Web is a huge repository of web pages and links. It provides abundance information for the Internet users. The growth of web is incredible as it can be seen in present days. Users’ accesses are recorded in web logs. Web usage mining is application of mining techniques in logs. The proposed system consists of two phases, (i) Offline phase and (ii) Online phase. In the offline phase, preprocessing and clustering is performed, while the classification and prediction is performed during the online phase. Preprocessing is the step which transforms the raw log file into a form that is more suitable for mining. Four steps are used in preprocessing, they are, data cleaning, user identification, session identification and formatting the result to suit the clustering algorithm. Modified ant-based clustering is proposed in this paper. Here, the algorithm doesn’t have any parameters and assumptions. The proposed method will automatically calculate the number of ants required for clustering. In the online phase, a classification algorithm based on Longest Common Sequence algorithm is used. The experimental results suggest that the proposed technique for web log mining results in better prediction of user behaviors when compared to the conventional techniques.