Temporal based factorization approach for solving drift and decay in sparse scoring matrix

Abstract

Collaborative filtering (CF) is one of the most popular techniques of the personalized recommendations, where CF generates personalized predictions in the rating matrix. The rating matrix typically contains a high percentage of unknown rating scores which is called the sparsity problem. The matrix factorization approach through temporal approaches has the accurate performance in addressing the sparsity issue but still with low accuracy. However, there are four issues when a factorization approach is adopted which are latent feedback learning, score overfitting, user’s interest drifting and item’s popularity decay over time. Therefore, this work introduces the temporal based factorization approach named TemporalMF++ to address all the issues. The experimental results show the TemporalMF++ approach has a higher prediction accuracy compared to the benchmark approaches. In summary, the TemporalMF++ approach has a superior effectiveness in improving the accuracy prediction of the CF by learning the temporal behaviour.