Improving Recommendation Performance with Clustering and Missing Value Prediction

High percentage of missing values in rating matrix results in low recommendation performance, which includes poor recommendation accuracy and slow recommendation generation. This paper proposes an integrated method with clustering algorithm and missing value prediction in a Collaborative Filtering Recommendation system to address the missing value issue in the rating matrix. The proposed method is validated using the MovieLens dataset. Experimental results show that the proposed method improves recommendation quality and online scalability while reducing recommendation generation time.