Research on Path Completion Technique in Web Usage Mining
Yan Li, Boqin Feng, Qinjiao Mao · 2008
An implementation of data preprocessing system for Web usage mining and the details of algorithm for path completion are presented. After user session identification, the missing pages in user access paths are appended by using the referer-based method which is an effective solution to the problems introduced by using proxy servers and local caching. The reference length of pages in complete path is modified by considering the average reference length of auxiliary pages which is estimated in advance through the maximal forward references and the reference length algorithms. As verified by practical Web access log, the proposed path completion algorithm efficiently appends the lost information and improves the reliability of access data for further Web usage mining calculations.