Integration of Clustering and Rule Induction Mining Framework for Evaluation of Web Usage Knowledge Discovery System

With the increasing popularity of WWW, some data such as users addresses or URLs which are being requested by the user are repeatedly collected by Web servers and collected in access log files. Investigating a server access data provides valuable information for performance improvement such as, reorganizing a web site for improving efficiency. Determining the path leading to accessed web pages which are often gathered into access log files is generally termed as web usage mining. In this connection, the concerned techniques mainly focus on the customer behavioral patterns discovered from a Web server log file in order to mine relationships within gathered data. The proposal in study, presents a novel framework, Integration of Clustering and Rule Induction Mining (ICRIM) which evaluate the performance of web usage knowledge discovery system. ICRIM framework incorporates the clustering model and Induction based decision rule model. The proposed evolutionary clustering model discovers web data clusters and analyzes the web site visitor behavior and optimally segregate similar user interests. Induction based decision rule model generates inferences and implicit hidden behavioral aspects in the web usage mining to investigate at the web server and client logs. Experimentation is carried out on ICRIM framework to evaluate the performance of web usage knowledge discovery system. Performance results of ICRIM framework are compared with existing clustering algorithm and induction based decision rule model.