Web Mining Through Customized Markov Algorithm

Abstract

The problem of predicting a user’s behavior on a web site has gained Importance due to the very increases required web network. It also compulsory for market and research of social behavior. Although many approaches are comes and many will should come, however they all are important at their different different view. Even I (Author, himself), suggests a physical approach to measure web traffic, results in paper titled “The Web Page Prediction by TRP Machine” , however I want to proposed an algorithm for it, which is not physically but logically measures the web traffic at all. Results as “Customized Page Rank algorithm” from simple Page Rank algorithm. Because simple Buddy Prima algorithm gives results after complicated calculations, instead it the novel approach creates less calculations and more clear and precise results as we want. The “Customized Buddy Prima algorithm” is based on the Markov Model with its chain in Forward reasoning [Research Paper on Forward Reasoning in Hierarchical representation , by me, (Author himself)], with the help of Apriori algorithm, Division algorithm, Association rule, page rank algorithm, Eigen vector calculations, modified Adjacency matrix, and many more calculations as tools. The suggested method is very proper as the result point of view, as it gives more proper results as shown by comparison of both two approaches as traditional one and the modified as well.