Fig. 3 | Scientific Reports

Fig. 3

From: Development and validation of a machine learning model for predicting mortality risk in veno-arterial extracorporeal membrane oxygenation patients

Fig. 3

ROC Curves for different machine learning models. (A) ROC curves for different machine learning models on the training dataset, with the x-axis representing the False Positive Rate (FPR) and the y-axis representing the True Positive Rate (TPR). Random Forest (AUC = 0.97) and CatBoost (AUC = 0.97) show the best performance, with Logistic Regression having the lowest AUC (0.79). (B) ROC curves for different machine learning models on the test dataset. Logistic Regression (AUC = 0.83) performs best, followed by Random Forest (AUC = 0.79) and DNN (AUC = 0.80). The models generally show good performance on the test data. (C) ROC curves for different machine learning models on the external validation dataset. DNN (AUC = 0.73) and SVM (AUC = 0.73) show the best performance, while Random Forest (AUC = 0.66) performs the worst. Overall, models tend to perform worse on the external validation data compared to the training and test datasets.

Back to article page