Abstract

Web Usage Mining is a computational process of discovering patterns in large data sets involving methods using the artificial intelligence, machine learning, statistical analysis and database systems with the goal to extract valuable information from accessing server logs of World Wide Web data repositories and transform it into an understandable structure for further understanding and use. Main focus of this paper will be centered on exploring methods that expedites the log mining process and present the result of log mining process through data visualization and compare data-mining algorithms. For the comparison between classification techniques, precision, recall and ROC area are the correct measures that are used to compare algorithms. Based on this study it shows that Naïve Bayes and Bayes Network are proven to be the best algorithms for that.

Authors and Affiliations