结合矩阵补全的宽度协同过滤推荐算法OA北大核心CSTPCD

Collaborative filtering is a classic method used in recommendation systems,designed to cater to the need for personalized recommendations.However,many collaborative filtering algorithms struggle when confronted with sparse rating data.To address this issue,we propose a broad collaborative filtering recommendation algorithm that integrates matrix completion.Initially,a matrix completion technique is employed to recover the user-item rating matrix.Sub-sequently,this completed rating matrix is utilized to identify respective neighbors for a given user-item pair,which in turn helps create the user-item rating collaboration vector.Finally,a broad learning system is employed to establish the complex nonlinear relationship between user-items and ratings.The effectiveness of the proposed algorithm has been validated through tests on MovieLens and Filmtrust data sets.The experimental results show that,compared with state-of-the-art collaborative filtering methods,the proposed method can effectively alleviate the data sparsity problem.It also possesses lower computational complexity and enhances recommendation performance to a certain extent.