Data Preprocessing Technique in Web Log Mining
Xiaofeng Lu · Jisuanji gongcheng · 2009
Data preprocessing is the important step in Web log mining. It consists of four sub-steps, i.e. data cleaning, user identification, session identification and path completion. The referer-based method is adopted for user session identification and path completion, in order to avoid the problems introduced by using proxy servers, firewall, local caching, and so on. Experimental results reveal that the technique can obtain the user access path efficiently if accurate referer information is available in Web access log.