Probabilistic Latent Factor Model for Collaborative Filtering with Bayesian Inference

Abstract

Latent Factor Model (LFM) is one of the most successful methods for Collaborative filtering (CF) in the recommendation system, in which both users and items are projected into a joint latent factor space. Base on matrix factorization applied usually in pattern recognition, LFM models user-item interactions as inner products of factor vectors of user and item in that space and can be efficiently solved by least square methods with optimal estimation. However, such optimal estimation methods are prone to overfitting due to the extreme sparsity of user-item interactions. In this paper, we propose a Bayesian treatment for LFM, named Bayesian Latent Factor Model (BLFM). Based on observed user-item interactions, we build a probabilistic factor model in which the regularization is introduced via placing prior constraint on latent factors, and the likelihood function is established over observations and parameters. Then we draw samples of latent factors from the posterior distribution with Variational Inference (VI) to predict expected value. We further make an extension to BLFM, called BLFMBias, incorporating user-dependent and item-dependent biases into the model for enhancing performance. Extensive experiments on the movie rating dataset show the effectiveness of our proposed models by compared with several strong baselines.

8 pages, 5 figures, ICPR2020 conference

References in corpus (8)

• Variational Inference: A Review for Statisticians
• Bayesian CP Factorization of Incomplete Tensors with Automatic Rank Determination
• Sequential Recommender Systems: Challenges, Progress and Prospects
• Research Commentary on Recommendations with Side Information: A Survey and Research Directions
• Maximum a Posteriori Estimators as a Limit of Bayes Estimators
• Local Popularity and Time in top-N Recommendation
• Collaborative Filtering with Label Consistent Restricted Boltzmann Machine
• The Epic Story of Maximum Likelihood