Personalized Recommendation Method Based on Web Log Mining

As Web log mining is belonged to one of major technologies and tools to discover user's interest, in this paper, we propose a novel personalized recommendation method based on Web log mining. The proposed personalized recommendation system contains offline and online module. We consider three types of Web log files in this paper, include: 1) Sever log, 2) Error log, and 3) Cookie log. In addition, we analyze the internal structure of the Web log file. The main innovation of this paper is to introduce collaborative filtering in personalized recommendation. Particularly, we assume that users with similar rating behaviors are possible to have similar interest to an item. Next, we utilize the hierarchical clustering technology to cluster users according to their profiles. Finally, experimental results demonstrate that the proposed algorithm is able to achieve higher personalized recommendation results and lower calculation time.