Recommender system based on latent topics

Abstract

Collaborative filtering is one of the most used techniques in recommender systems. The goal of this paper is to propose a new method that uses latent topics to model the items to be recommended. In this way, the ability to establish a similarity between these elements is incorporated, improving the performance of the recommendation made. The performance of the proposed method has been measured in two very different contexts, yielding satisfactory results. Finally, the conclusions and some future lines of work are included.