Analysis Of Web Users Based On The Fuzzy Clustering

Web usage mining has been a hotspot in the field of data mining. Web usage mining is a process to find out the usage patterns of Web from the Web usage data or the log files with the data mining techniques. Web usage mining has three phases: Web data preprocessing, patterns discovery, and patterns analysis.The paper analyses the definitions of Web usage mining and clustering, their corresponding techniques and current research status. Finally, we present a scheme to analyzing the Web users based on the web usage mining.The essence idea of the Web users analysis based on the web usage mining is: analyzing the Web log data, using the data mining approaches to find out the users pattern, and providing useful users' information for web making decision. Because the uncertainty of the user segments we use the fuzzy k-means cluster algorithm; to overcome the dependence on the centers selection we use the hierarchical method to get the centers of k cluster first, which decreases this dependence. We also discuss the increment clustering techniques which can improve the efficiency of cluster analysis with the prophase results to deal with the increment data. The experiment indicates that the new approaches have improved the astringency of the fuzzy k-means.