Combining the paths similarity and the pages content similarity, a novel clustering algorithm is presented. The actions character
of users is revealed more exactly by clustering. The data scale is reduced by a long way during clustering. Based on the clusters,
the user navigation patterns are generated by mining the Web log. The experiment result shows that the user navigation interest
conversion patterns mined from Web log are typical and intuitionistic.