A novel constrained non-negative matrix factorization method based on users and items pairwise relationship for recommender systems
作者：
Highlights：
• NMF generates low-rank data while keeping the non-negativity of matrices elements.
• Update rules warrant the non-negativity of factorized matrices and the convergence.
• Constrained NMF imposes constraints as regularizers in the factorizing process.
• CNMF uses the Frobenius norm and the KL-divergence as the cost functions.
• CNMF applies penalties to create a generic and flexible algorithm for RSs.
摘要
•NMF generates low-rank data while keeping the non-negativity of matrices elements.•Update rules warrant the non-negativity of factorized matrices and the convergence.•Constrained NMF imposes constraints as regularizers in the factorizing process.•CNMF uses the Frobenius norm and the KL-divergence as the cost functions.•CNMF applies penalties to create a generic and flexible algorithm for RSs.
论文关键词：Recommender systems,Non-negative matrix factorization,Multiplicative updating rules,Latent factors
论文评审过程：Received 7 June 2020, Revised 14 January 2022, Accepted 19 January 2022, Available online 8 February 2022, Version of Record 10 February 2022.
论文官网地址：https://doi.org/10.1016/j.eswa.2022.116593
