Collaborative filtering using non-negative matrix factorisation

Collaborative filtering is a popular strategy in recommender systems area. This approach gathers users’ ratings and then predicts what users will rate based on their similarity to other users. However, most of the collaborative filtering methods have faced problems such as sparseness and scalability. This paper presents a non-negative matrix factorisation method to alleviate these problems via decomposing rating matrix into user matrix and item matrix. This method tries to find two non-negative user matrix and item matrix whose product can well estimate the rating matrix. This approach proposes updated rules to learn the latent factors for factorising the rating matrix. The proposed method can estimate all the unknown ratings and its computational complexity is very low. Empirical studies on benchmark datasets show that the proposed method is more tolerant of the sparseness and scalability problems.

[1]  Gediminas Adomavicius,et al.  Stability of Recommendation Algorithms , 2012, TOIS.

[2]  John Riedl,et al.  Item-based collaborative filtering recommendation algorithms , 2001, WWW '01.

[3]  Yehuda Koren,et al.  Matrix Factorization Techniques for Recommender Systems , 2009, Computer.

[4]  Peyman Kabiri,et al.  Modelling trust networks using resistive circuits for trust-aware recommender systems , 2017, J. Inf. Sci..

[5]  Zhen Lin,et al.  Incorporating social information to perform diverse replier recommendation in question and answer communities , 2016, J. Inf. Sci..

[6]  Gang Chen,et al.  Collaborative Filtering Using Orthogonal Nonnegative Matrix Tri-factorization , 2007, Seventh IEEE International Conference on Data Mining Workshops (ICDMW 2007).

[7]  Thomas Hofmann,et al.  Latent semantic models for collaborative filtering , 2004, TOIS.

[8]  Hamed Movahedian,et al.  Folksonomy-based user interest and disinterest profiling for improved recommendations: An ontological approach , 2014, J. Inf. Sci..

[9]  Jun Wang,et al.  Unifying user-based and item-based collaborative filtering approaches by similarity fusion , 2006, SIGIR.

[10]  Yehuda Koren,et al.  Factor in the neighbors: Scalable and accurate collaborative filtering , 2010, TKDD.

[11]  Jonathan L. Herlocker,et al.  Evaluating collaborative filtering recommender systems , 2004, TOIS.

[12]  Reza Rafeh,et al.  An adaptive approach to dealing with unstable behaviour of users in collaborative filtering systems , 2012, J. Inf. Sci..

[13]  Lior Rokach,et al.  Introduction to Recommender Systems Handbook , 2011, Recommender Systems Handbook.

[14]  M. Analoui,et al.  Application of nonnegative matrix factorization in recommender systems , 2012, 6th International Symposium on Telecommunications (IST).

[15]  Qiang Yang,et al.  Scalable collaborative filtering using cluster-based smoothing , 2005, SIGIR '05.

[16]  Andrzej Cichocki,et al.  Nonnegative Matrix and Tensor Factorization T , 2007 .