A Comparative Analysis on Markov Model (MM) Association Rule Mining (ARM), Association Rule Mining-Statistical Features (ARM-SF), Boosting and Bagging Model (BBM) to Impervious Web Page Prediction

Abstract

The web usage mining techniques are used to analyze the web usage patterns for a web site. To fetch the user access patterns from the user log files. The access patterns are used in the prediction process. Web prediction is a classification model attempts to predict the next set of Web pages that a user may visit based on the knowledge of the previously visited pages. By Predicting user’s behavior can be applied effectively in various critical applications in the internet environment.So that the application has traditional tradeoffs between modeling complexity and prediction accuracy. The association rule mining and classification techniques are used to perform the pattern extraction and prediction process. The technique as Markov model and all Kth Markov model are used in Web prediction. A modified Markov model is proposed to alleviate the issue of scalability in the number of paths. The framework can improve the prediction time without compromising prediction accuracy. The proposed system integrates the boosting and bagging models with the association rule mining technique to improve the prediction accuracy. The statistical features are also discovered and used in the proposed model. Standard benchmark data sets are used to analyze the web page prediction process. The proposed model reduces the number of paths without compromising accuracy. The system improves the accuracy with scalability considerations.