Services of prediction for visiting path based on improved matrix clustering

According to the analysis of Web log files, matrix clustering algorithm not only can be optimized, but also mining results can be applied to predicting users´ visiting path and classify new users. The data-preprocessing phase has been discussed in detail. Then, Web log mining algorithms and pattern analysis and appliance phase are presented, including the optimized matrix clustering algorithm and the basic idea and realization of the prediction algorithm. Experiments show that the algorithm is effective and feasible.