Fair performance-based user recommendation in eCoaching systems

Abstract Offering timely support to users in eCoaching systems is a key factor keep them engaged. However, coaches usually follow lot of users, so it hard for prioritize those with whom they should interact first. Timeliness especially needed when health implications might be the consequence lack support. In this paper, we focus on last scenario, by considering an platform runners. Our goal provide coach ranked list according need. Moreover, want guarantee fair exposure ranking, make sure that different groups have equal opportunities get supported. order do so, first model their performance and running behavior then present ranking algorithm recommend coaches, session quality previous ones. We measures fairness allow us assess propose re-ranking exposure. Experiments data coming from previously mentioned runners show effectiveness our approach standard metrics assessment its capability users. The source code preprocessed datasets are available at: https://github.com/wiguider/Fair-Performance-based-User-Recommendation-in-eCoaching-Systems .