Skip to main content
. 2021 Jan 14;11:1378. doi: 10.1038/s41598-021-80998-y

Figure 5.

Figure 5

ROC curves of models using four feature reduction/fusion methods. (A) ROC curve for PCA based fusion method, AUC = 0.60, specificity = 0.58, sensitivity = 0.64. (B) ROC curve for Boruta based feature reduction method, AUC = 0.60, specificity = 0.47, sensitivity = 0.48. (C) ROC curve for CPH based feature reduction method, AUC = 0.55, specificity = 1.00, sensitivity = 0.18. (D) ROC curve for LASSO based feature selection method, AUC = 0.50, specificity = 0.26, sensitivity = 0.91. (E) ROC curve for the proposed risk-score based feature fusion method, AUC = 0.84, specificity = 0.68, sensitivity = 0.91.