Collaborative Filtering Hybrid Recommendation Algorithm Based on Improved Score Similarity

Collaborative filtering recommendation algorithm is one of the most widely used recommendation algorithms in the recommendation system, but the traditional similarity calculation method only considers the user-item rating data, and there are still great defects in the face of data sparsity and cold start, which leads to inaccurate similarity calculation results and low accuracy of recommendation results. In order to solve the above problems, this paper proposes a collaborative filtering hybrid recommendation algorithm based on improved rating similarity (ISSH-CF). In the calculation of traditional similarity, the hot item penalty factor and the active user penalty factor are introduced, and the similarity of user and item attribute characteristics is combined respectively, which alleviates the user cold start problem, the item cold start problem and the data sparsity problem. For old users, finally, the rating prediction of UserCF and ItemCF were calculated respectively, and the weighted hybrid recommendation was performed to obtain the final rating prediction result. Compared with the original UserCF and ItemCF, the MAE values are reduced by 2.81%and 2.78%respectively on the MovieLens 100k data set, which proves that the proposed algorithm can effectively improve the accuracy of recommendation results.