The personality-aware recommender system by [Lu and Tintarev, 2018] creates recommendations for a user by weighting items via a linear combination of rank (predicted by a factorization machine) and diversity, the latter leveraging a correlation model between OCEAN scores and diversity preferences.

The BFI-44 questionnaire commonly provides a more accurate estimate of an individual's personality traits in comparison to the TIPI questionnaire.

The location/point-of-interest recommender proposed by [Ravi and Vairavasundaram, 2017] adopts a deep graph-embedding method, into which emotion information extracted from microblogs are infused via a gating mechanism.

Affect can be modeled via categorical, dimensional, and regressional models.

Important reasons to incorporate personality information into a recommender system include mitigating cold start, alleviating popularity bias, and diversifying recommendations.

Consider the following preferences and needs of a user who is well versed in technical aspects of recommender systems. Given a new recommender system, the user wants to evaluate it. Help him or her choose the most appropriate evaluation measure/metric to quantify the described property of the system.