A personalized collaborative recommendation algorithm based on user clustering smoothing

The personalized collaborative recommendation algorithm has been successfully applied to various electronic commerce recommender systems. Unfortunately, with the tremendous growth in the amount of items and users, the lack of original rating poses some key challenges for recommendation quality. And the recommendation systems suffered from its shortage in scalability as their calculation complexity and space complexity increased quickly. To solve the scalability problem in the personalized recommendation systems, the paper proposed a personalized collaborative recommendation algorithm based on user clustering smoothing. The algorithm consists of five steps of clustering the users based on k means algorithm, smoothing the vacant ratings where necessary, selecting the user clustering centers, forming neighbors from the selected user centers, and at last producing recommendations. The personalized collaborative algorithm based on user clustering smoothing can alleviate the scalability problem in the recommender systems.