Implementation Of Web Usage Mining Using Log Markup Language (LOGML) and Closed Association Rule Mining Algorithm (CHARM)

Abstract

Web usage mining is the discovery of user access patterns from Web usage logs. The paper presents two XML (Extensible Markup Language) 1.0 applications and a web data mining application which utilizes it to extract web data from web log files. The two XML 1.0 applications are: LOGML (Log Markup Language) is a web-log report description language and XGMML (Extensible Graph Markup and Modeling Language) is a graph description language. As a case study, the system implements a sample website for web graph information and usage information. The main goal of this paper is that web graph information transforms to XGMML document and then this graph information and cleaned usage information of finished user sessions transforms to LOGML document. The next goal is that these cleaned data with LOGML document are mined with CHARM (Closed Association Rule Mining) algorithm to implement the most frequently accessed pages from one site. CHARM is an efficient algorithm for mining all closed frequent itemsets (set of all subsets of items).