Thilagu M and Nadarajan R
PSG College of Technology, India

In recent times, web usage mining has becomea promising solution for discovering interesting patterns from web data. The discovered knowledge is useful for web developersor designerstounderstandusers’ navigation behaviour and help inimprovingwebsite design, performanceand e-business strategies.Still, analyzingthe dynamic behaviour of users’ intereston web traversal patternsfrom the underlying web usage datais an important task in order toanalyze the significantchanges in users’ access to web pages and time spent onthem.In this paper, an investigation onusers’ access behaviour of web traversal patterns in terms of support andutilityis doneand discussed.Algorithmshave been proposed to discover significant web traversal patterns based on support andutility constraints in Phase-I. And,theproposed algorithmsapply transitional pattern mining modelsto detectsignificant milestones or time points at which users’ intereston the discovered patternsincreases or decreases dramatically, in Phase-II. The information obtained from the proposed algorithmshelps in understanding users’ dynamic preferencesof web traversal patterns. Experimental studies are done on a real dataset and the results are discussed

Web Usage Mining, Significant Web Traversal Pattern, Significant Milestone, Frequent Pattern, High Utility Pattern