Nowadays, the capability to generate and collect data has been expanded enormously. Data mining, which is the process of extracting useful information from large databases, has been applied in many business areas for making business decisions.
With the rapid progress of World Wide Web technology and its ever-growing popularity, a great number of logfiles, which are web page access histories, are being collected. Logfiles contain such useful information as client IP address, access time, request method, requested file name, error code, protocol version, browser and OS.
This thesis considers web mining by applying data mining techniques for analyzing logfiles which are generated using Common Logfile Format. Association rule analysis methods are developed, and error analysis and referrer analysis are also conducted. The results can be used to provide some strategies for web server performance enhancement, web site design improvement and effective web site advertisement.
