Clustering and Correlation based Collaborative Filtering Algorithm for Cloud Platform

With the development of the Internet, recommender systems have played a more and more important role in the field of big data processing, such as e-business. In order to deal with big data in recommender systems, we propose a clustering and correlation based collaborative filtering algorithm for cloud platform, which improves the traditional user-based collaborative filtering algorithm with k-medoids clustering and a data structure named correlation multi-tree in this paper. Firstly, we analyze the user-based collaborative filtering for cloud platform. On the basic of it, we propose a k-medoids based collaborative filtering algorithm for cloud platform by using the k-medoids clustering. It can solve the problem of data sparsity effectively. As a result, it can be more efficient with the recall rate and recommendation ratings improved at the same time. Considering the falling of recommendation accuracy by using clustering technology, this paper introduces a data structure named correlation multi-tree to correlate the user information and their neighbors information. It can be used to compute the extended user-item score, which makes full use of the correlation between data on cloud platform. As a result, the clustering and correlation based collaborative filtering algorithm for cloud platform proposed by us can improve the recommendation accuracy effectively, and ensure the effect of recommendation and the time efficiency at the same time. An extensive experimental evolution with Ali data sets on Hadoop cloud platform shows that our collaborative filtering algorithm has a better recommendation and is more efficient in handling big data.