Insights into Imaging

Table 5 Comparison of ROC curves among different models by DeLong test

From: Feasibility and effectiveness of automatic deep learning network and radiomics models for differentiating tumor stroma ratio in pancreatic ductal adenocarcinoma

KNN/SVM	0.8313	SVM/LR	0.8927	LR/SN	0.3350	RF/Res	0.0206*
KNN/LR	0.7372	SVM/RF	0.8394	LR/Xec	0.0196*	SN/Xec	0.2722
KNN/RF	0.5784	SVM/SN	0.3140	LR/Mob	0.0276*	SN/Mob	0.2742
KNN/SN	0.1656	SVM/Xec	0.0534	LR/Res	0.0165*	SN/Res	0.1053
KNN/Xec	0.0106*	SVM/Mob	0.0368*	RF/SN	0.3752	Xec/Mob	0.9100
KNN/Mob	0.0051*	SVM/Res	0.0243*	RF/Xec	0.0238*	Xec/Res	0.6912
KNN/Res	0.0054*	LR/RF	0.9430	RF/Mob	0.0379*	Mob/Res	0.8273

KNN knearest neighbor, SVM support vector machine, RF random forest, LR logistic regression, SN ShuffulNetV2, Xec Xception, Mob MobileNetV3, Res ResNet18
^*Represents p < 0.05

Back to article page