Advancing Loss Functions in Recommender Systems: A Comparative Study with a Rényi Divergence-Based Solution