Performance Evaluation of Clustering in Web-Log Analysis Based on Agent

Web mining is the use of data mining Technique to automatically discover & extract information from web documents. When user searches for goods the management agent receives order from graphical user interface .Management agent receives information, update agent information store house and feedback the mining result to user. Intelligent agent can help making computer system easier to use, enable finding & filtering information. The mining agent is the analytical center of whole agent system.It mainly adopts two kind of analytical method: related rule mining and cluster analysis. Cluster of objects are formed so that objects with in a cluster have high similarity. The aim of this paper is to analyze the web log data .To achieve this clustering tool is used. It performs in two phases. First it captures the web-log data. Then it analyzes the data& discovers the hidden pattern. Agent requires an agent communication language to describe & process agent request. The future internet will use PERL to encode information with meaningful structure & semantics.