SOM Improved Neural Network Approach for Next Page Prediction

The increasing usage of web results the heavy communication and slow returns from web. Because of this, there is the requirement of some approaches to optimize the web resources usage. One of such approach is caching that can be used within an organization to optimize the access of frequently used web pages. Caching is about to predict the requirement of next web access of a user and load it in cache before user request. This kind of intelligent prediction comes under web usage mining. In this work, an intelligent SOM(Self Organizing Map) improved neural network approach is defined to perform next web page prediction. The work will be here presented in three main stages. In first stage, to perform the intelligent sequence mining the dataset will be filtered. The filtration will be here performed using clustering approach. The clustering will be performed based on web usage. Now only the cluster that represents the high usages pages will be considered for prediction. In second stage, SOM will be applied to analyze the web page usage and the prediction of next required page access. The SOM will be applied here to assign the weightage to next possible based on frequency and time stamp analysis. Once the weightage will be applied, the final work is to apply the neural network to predict the next visiting page. GENERAL TERMSWeb Caching, Web prefetching, web usage mining KEYWORDSSOM, WWW, HTML