Study on dual K-means algorithm in collaborative filtering system based on web log

A collaborative filtering algorithm is designed and improved in this paper, which can optimize and improve the sparsity and extention of data in systems. We can Use the user's implicit information in the web log to cluster analysis on user set and by using the dual K-means algorithm, and we can use the K-means algorithm in each clustering set to further improve the effectiveness of the recommendation of the target users. By contrasting the experimental results, this method compared to the performance of other collaborative filtering algorithms have better performance in those sites browsed by large amount of users.