In their newly published article ‘Fairness Metrics and Bias Mitigation Strategies for Rating Predictions’ in the journal Information Processing & Management, Dr. Christian Haas and his co-author investigate algorithmic fairness research in the recommender system domain. Biases and questions of fairness can arise in automated decisions, and while much of the current research has focused on scenarios in machine learning and natural language processing, unfairness and the propagation of biases can also occur in recommender systems. The article explores similarities between bias and fairness research in machine learning and recommender systems, and develops a novel bias mitigation strategy to improve fairness in rating predictions.

Article DOI: https://doi.org/10.1016/j.ipm.2021.102646