Fig. 3From: A machine-learning approach to predict postprandial hypoglycemiaFlowchart of the proposed approach including data-preprocessing, and how to train RF, SVM-LN, SVM-RBF, KNN, and LR. Since the 5-fold cross-subject validation was used, training and testing models were repeated by 5 times. For each iteration, each model’s result and tuned hyper-parameters was savedBack to article page