Skip to main content
. 2022 Apr 7;12:852809. doi: 10.3389/fonc.2022.852809

Figure 3.

Figure 3

ROC curves in delayed phase. The yellow curve is the poor prognosis group (PP group), and the blue curve is the good prognosis group (GP group). (A) ROC curve of the KNN model in the validation set. The AUC were 0.462 in the PP group (sensitivity and specificity were 0.62 and 0.42, respectively) and 0.462 in the GP group (sensitivity and specificity were 0.42 and 0.62, respectively). (B) ROC curve of the SVM model in the validation set. The AUC were 0.532 in the PP group (sensitivity and specificity were 0.54 and 0.50) and 0.532 in the GP group (sensitivity and specificity were 0.50 and 0.54, respectively). (C) ROC curve of the XGBoost model in the validation set. The AUC were 0.609 in the PP group (sensitivity and specificity were 0.62 and 0.50, respectively) and 0.609 in the GP group (sensitivity and specificity were 0.50 and 0.62, respectively). (D) ROC curve of the RF model in the validation set. The AUC were 0.564 in the PP group (sensitivity and specificity were 0.54 and 0.58, respectively) and 0.564 in the GP group (sensitivity and specificity were 0.58 and 0.54, respectively). (E) ROC curve of the LR model in the validation set. The AUC were 0.628 in the PP group (sensitivity and specificity were 0.62 and 0.67, respectively) 0.578 in the GP group (sensitivity and specificity were 0.67 and 0.62, respectively). (F) ROC curve of the DT model in the validation set. The AUC were 0.359 in the PP group (sensitivity and specificity were 0.38 and 0.33, respectively) and 0.528 in the GP group (sensitivity and specificity were 0.33 and 0.38, respectively).