PREDICTION OF NEXT ACCESSED WEB PAGE USING CLOSPAN ON DHMM

Abstract

Web page prediction is based on user navigation patterns which are retrieved from the web logs of servers from various places that acts as standard for the prediction. The disadvantage of markov model was low coverage which was corrected by developing dynamic markov model. Also the orders of the markov model are defined based on the levels specified. The drawback of DHMM was high execution times and redundant links between the nodes which consumes times in processing. The proposed paper consist of new web page prediction based on CloSpan on DHMM. The graphs are plotted to show the comparison and the time execution difference between the base and the new proposed algorithm. The results obtained are far better than the base.