Time Domain based Web Usage Mining

A. Erdem (Japan) and A. Gursoy (Turkey)


Web Usage Mining, Web Site Improvement, Sequence Mining, Web Mining


Many web usage mining applications use frequent for ward references for characterizing user navigation behaviour. However, interpretation of forward and back ward references together with time domain can provide more precise information about interests of user. Relative time spent for each page in a session is considered as interest criterion. Since frequent web sequences may cover redundant path trials previous to reaching interested pages due to misleading guidance, such redundant frequent patterns should be distinguished from user interests. In this paper, we suggest incorporating time do main information into temporal knowledge that covers both forward and backward sequences for understanding user interests and non-interests. We used this knowledge for detecting web site deļ¬ciencies and used for web site improvement suggestions.

Important Links:

Go Back