Table 3.

Summary results for models trained and tested on each population (control or amputee). Sensitivity and specificity values represent the optimal point of the ROC curve.

Method and performance metric		Model, mean (1.96 SEM)
		Control-control	Control-amputee	Amputee-amputee
Stacking
	AUC	0.997 (0.003)	0.996 (0.004)	0.995 (0.004)
	Sensitivity	0.979 (0.022)	0.989 (0.017)	0.984 (0.016)
	Specificity	0.991 (0.012)	0.968 (0.025)	0.965 (0.022)
Threshold-based
	AUC	0.960 (0.020)	0.939 (0.059)	0.939 (0.059)
	Sensitivity	0.915 (0.040)	0.878 (0.097)	0.878 (0.097)
	Specificity	0.927 (0.045)	0.922 (0.045)	0.922 (0.045)