Key points are not available for this paper at this time.
In this paper, we first define the problem of item-ranking promotion (IRP) in recommender systems as (Goal 1) maintaining a high level of overall recommendation accuracy while (Goal 2) recommending the items with extra values (i.e., RP-items) to as many users as possible. Our novel framework, proposed to address the IRP problem, is based on our own loss function that simultaneously aims to achieve the two goals above and employs a learning-to-rank scheme for training a recommender model. Via extensive experiments, we validate the effectiveness of our framework in terms of the exposure rate of RP-items and the accuracy of recommendation.
Bae et al. (Sun,) studied this question.