A listwise collaborative filtering based on Plackett-Luce model

In recent years, ranking-oriented collaborative filtering (CF) algorithms have achieved great success in recommender systems. They achieve advanced performance by predicting item preference ranking rather than the absolute value of the item. However, the listwise collaborative filtering (ListCF) only considers the impact of user ratings, ignores the influence of other feature factors, which leads to lower accuracy of recommendation. This paper proposes a listwise collaborative filtering algorithm based on user ratings and user-item type ratings. Experiments on Movielens proved the improvement of recommendation accuracy.