Objective Evaluation or Subjective Evaluation in Digital Social Media
Zhenfei Feng, Laurence Favier · 2018
Recommender systems exploit users' ratings to infer their preferences in order to propose recommendations of items that match up with users' interests. Recently, online movie review sites have combined RS and social media. However, in these sites, ratings can have different meaning for users: users may choose one criterion or another. Thus, users' ratings might not be reliable enough to infer their preference in this context. In this study, we find that users may rate movie either by referring to perceived movie quality and/or by referring to their own preferences. Furthermore, users chose different criteria in different contexts. We discuss how users' ratings can be interpreted more precisely in order to improve the performance of recommender systems.