Combining Singular Value Decomposition and Item-based Recommender in Collaborative Filtering

Recommender Systems are introduced as an intelligent technique to deal with the problem of information and product overload. Their purpose is to provide efficient personalized solutions in economic business domains. Collaborative filtering is a widely used method of providing recommendations using ratings on items from users. However, it has three major limitations, accuracy, data sparsity and scalability. This paper proposes a new collaborative filtering algorithm to solve the problems mentioned above. We utilize the results of singular value decomposition (SVD) to fill the vacant ratings and then use the item based method to produce the prediction of unrated items. Our experimental results on MovieLens dataset show that the algorithm combined SVD method and item-based method is promising, since it does not only solute some of the recorded problems of recommender systems, but also assists in increasing the accuracy of systems employing it.