Multi-Objective Optimization Recommendation Algorithm Based on Collaborative Filtering and Item Similarity
Abstract— The traditional collaborative filtering recommendation algorithm generates recommendations for users by calculating the similarity between users or items, without considering other factors and data sparsity problem. The appearance of personalized tags can reflect the characteristics of items. Some algorithms try to optimize the recommendation results by setting weight coefficients of various factors. However, it is unreasonable to set the same weight coefficients for different users. To solve these problems, we propose a multi-objective optimization recommendation algorithm based on collaborative filtering and item similarity. First. The algorithm calculates the similarity between users by improved user-based collaborative filtering recommendation algorithm and generates a candidate recommendation set. Then, the algorithm uses the coupled object similarity measure method to calculate the similarity between items. Finally, we use NNIA algorithm to optimize the former two goals and generate the final recommendation results. Experimental results on the Movielens dataset shows that the proposed algorithm has good recommendation effect.
Index Terms— collaborative filtering, item similarity, multi-objective optimization, NNIA.
School of Big Data & Software Engineering, Chongqing University, CHINA
Cite: Chaosheng Zhao, "Multi-Objective Optimization Recommendation Algorithm Based on Collaborative Filtering and Item Similarity," Proceedings of 2019 the 9th International Workshop on Computer Science and Engineering, pp. 99-107, Hong Kong, 15-17 June, 2019.