The role of user characteristics in the evaluation of recommender systems recommendations

What is it about?

Recommender Systems are often used in domains with diverse users with different needs and requirements. Explanations can help these users understand why they see the recommendations or make better decisions. However, the same explanation might have different effects on different users. This paper examines the various user types on which explanations were evaluated and how the measured impact of explanations differs between the users.

Featured Image

Why is it important?

Our findings reveal research gaps that should be addressed by the community. First, the users on which the explanations are evaluated often do not reflect the diversity of the target users of the domain in which the recommender system is operating. Second, when the effect of an explanation is evaluated, it is rarely analyzed how the effect differs between users with varying characteristics.