Machine learning-based forecast of Helmet-CPAP therapy failure in Acute Respiratory Distress Syndrome patients

Riccardo Campi, Antonio De Santis, Paolo Colombo, Paolo Scarpazza, Marco Masseroli

Abstract: Highlights•Predicting the H-CPAP therapy failure for ARDS patients using Machine Learning.•Using hospital EMRs to train a SVM Classifier and a Neural Network Classifier.•Altered levels of PaO2/FiO2 Ratio and CRP can cause H-CPAP failure in ARDS patients.•Using a feature reduction algorithm to minimize the required number of features.