Improving Web Page Prediction Using Default Rule Selection

Abstract

Mining user patterns of web log files can provide significant and useful informative knowledge. A large amount of research has been done in trying to predict correctly the pages a user will most likely request next. Markov models are the most commonly used approaches for this type of web access prediction. Web page prediction requires the development of models that can predict a user’s next access to a web server. Many researchers have proposed a novel approach that integrates Markov models, association rules and clustering in web site access predictability. The low order Markov models provide higher coverage, but these are couched in ambiguous rules. In this paper, we introduce the use of default rule in resolving web access ambiguous predictions. This method could provide better prediction than using the individual traditional models. The results have shown that the default rule increases the accuracy and model-accuracy of web page access predictions. It also applies to association rules and the other combined models.