A novel missing data imputation approach based on clinical conditional Generative Adversarial Networks applied to EHR datasets

Michele Bernardini, Anastasiia Doinychko, Luca Romeo, Emanuele Frontoni, Massih-Reza Amini

Abstract: Highlights•We propose a novel clinical conditional Generative Adversarial Network (ccGAN).•ccGAN imputes missing values from multi-diabetic centers’ routine EHR data.•ccGAN exploits fully-available features to infer other missing clinical features.•ccGAN overcomes significantly other state-of-the-art imputation methodologies.•ccGAN is integrated into a clinical decision support system for screening purposes.