A collaborative filtering recommendation algorithm based on user clustering and Slope One scheme

Recommendation system has been widely used in electronic commerce, news, web2.0, E-learning and other fields. Collaborative filtering is one of the most important algorithms. But as scale of recommendation system continues to expand, more and more problems appear. Data sparsity and poor prediction are main problems that recommendation system has to face. To improve the quality and performance, a new collaborative filtering recommendation algorithm combining user-clustering and Slope One algorithm is proposed. In our algorithm, users were clustered into several classes based on users' rating on items; therefore the useless information was filtered. Then the slope-one scheme was applied to predict the object rating. The experiments were applied to the MovieLens dataset to exploit the benefits of our detector and the experiment results show that the accuracy of our algorithm is in advance of previous research.