Optimal machine learning methods for prediction of high-flow nasal cannula outcomes using image features from electrical impedance tomography

The xgboost method showed better overall performance for balanced EIT image features, which may be considered as the ideal machine learning method for early prediction of HFNC outcomes.