The Magic Barrier of Recommender Systems – No Magic, Just Ratings

Abstract

Abstract. Recommender Systems need to deal with different types of users who represent their preferences in various ways. This difference in user behaviour has a deep impact on the final performance of the recommender system, where some users may receive either better or worse recommendations depending, mostly, on the quantity and the quality of the information the system knows about the user. Specifically, the inconsistencies of the user impose a lower bound on the error the system may achieve when predicting ratings for that particular user. In this work, we analyse how the consistency of user ratings (coherence) may predict the performance of recommendation methods. More specifically, our results show that our definition of coherence is correlated with the so-called magic barrier of recommender systems, and thus, it could be used to discriminate between easy users (those with a low magic barrier) and difficult ones (those with a high magic barrier). We report experiments where the rating prediction error for the more coherent users is lower than that of the less coherent ones. We further validate these results by using a public dataset, where the magic barrier is not available, in which we obtain similar performance improvements.

Can Matrix Factorization Improve the Accuracy of Recommendations Provided to Grey Sheep Users?

“…These conclusions are consistent with those reported in (Bellogín et al, 2014), that showed that with a neighbor-based approach, exploiting only inconsistent users to compute recommendations for inconsistent users does not lead to accurate recommendations.…”

“…This is also the conclusion reported in (Bellogín et al, 2014): "less coherent users need information from outside of their own cluster".…”

“…They refer to these users whose opinions do only partially agree or disagree with any community of users (Claypool et al, 1999). Inconsistent users (Bellogín et al, 2014) may also be related to GSU. Inconsistency is defined as users' inherent noise when interacting (rating) with a recommender system.…”

“…The accuracy of the recommendations provided to individual users is slightly improved. (Bellogín et al, 2014) have studied the accuracy of the recommendations that inconsistent users receive. The entire set of users is also split into easy users (normal users) and difficult (inconsistent) users.…”

A Novel Framework to Process the Quantity and Quality of User Behavior Data in Recommender Systems

“…Several approaches have been introduced to handle the natural noise in recommender systems [3,5,11,13]. O'Mahony et al [13] classified noise in recommender systems into natural (naturally occurred) and malicious (deliberately inserted) and proposed to remove noisy ratings to improve accuracy.…”