Title: Weighted-SVD: Matrix Factorization with Weights on the Latent Factors

Abstract: The Matrix Factorization models, sometimes called the latent factor models, are a family of methods in the recommender system research area to (1) generate the latent factors for the users and the items and (2) predict users' ratings on items based on their latent factors. However, current Matrix Factorization models presume that all the latent factors are equally weighted, which may not always be a reasonable assumption in practice. In this paper, we propose a new model, called Weighted-SVD, to integrate the linear regression model with the SVD model such that each latent factor accompanies with a corresponding weight parameter. This mechanism allows the latent factors have different weights to influence the final ratings. The complexity of the Weighted-SVD model is slightly larger than the SVD model but much smaller than the SVD++ model. We compared the Weighted-SVD model with several latent factor models on five public datasets based on the Root-Mean-Squared-Errors (RMSEs). The results show that the Weighted-SVD model outperforms the baseline methods in all the experimental datasets under almost all settings.