User Behavior Prediction Using Enhanced Pattern Tree Data Structure and Web Usage Mining

As we all know that, in today's era World Wide Web places a measurable role by providing lots of information to the user which will be more useful. Thus, to know the process of discovering and analyzing usage patterns of user, the paper aims to predict the behavior of the user from the web logs. The weblogs is termed as web log mining or web usage mining, every click made by the user will be automatically entered into the weblogs of the corresponding web server. The proposed data structure, pattern tree can be used efficiently to store the usage patterns and their frequency of all the users using the path sharing. The Enhanced Pattern Tree (EPT) maintains the relationship between different patterns and the corresponding users as rules. Search for the specific pattern would yield the corresponding user and vice-versa in minimum no. of searches as rule sharing is used by the pattern tree data structure. This research work aims to develop a framework for analyzing user behaviour through user patterns obtained from the web server logs and supports the use of association rules for representing the relationship between user and patterns.