Date de l'évènement :
à 12:00, (Paris, Europe)
Lieu de l'évènement :
En ligne
Professeur Roch GIORGI - Equipe QuanTIM-SESSTIM
Intervenant :
Richard RILEY
Institute of Applied Health Research, University of Birmingham, UK
Description :
Clinical prediction models estimate an individual’s risk of a particular health outcome. In this talk, I suggest that prediction model research has become the academic’s playground, and that machine learning and AI have only exacerbated this problem. I highlight common pitfalls, including model instability and poor calibration, and propose some solutions moving forwards. The talk is aimed at a broad audience.
Keywords: Clinical prediction models, machine learning, AI, predictive algorithms, stability
