Web mining for better web usability

Abstract

To develop more effective user-oriented learning techniques for the Web, we need to be able to identify a meaningful session unit from which we can learn. Without this, we could have a high risk to mix up the different user’s activity in the web. We are interested to detect boundaries of sequences between related sessions that would group the activities for a learning purpose. But identification of user session is not always easy where logged on and cookie information is not available. The problem of predicting user access in web pages has recently gets a significant attention. Several algorithms have been proposed, which find important applications, like user profiling, web perfecting, design of adaptive web sites, etc. In all these applications the main issue is the development of an effective prediction system. Because of its importance in reducing user perceived latency present in every Web-based application, which is a usability issue. This thesis paper describes a data mining technique for identify user sessions from huge amount of web log data and a web system, which makes prediction about the user target page by using those sessions to guide the user in World Wide Web.