Table 3.

Algorithm performance (development and validation set)

	Development set			Validation set
	ROC–AUC (95% CI)	Sensitivity (95% CI)	Specificity (95% CI)	ROC–AUC (95% CI)	Sensitivity (95% CI)	Specificity (95% CI)
KNN	0.923 (0.907 to 0.937)	0.856 (0.792 to 0.910)	0.850 (0.827 to 0.871)	0.925 (0.904 to 0.941)	0.891 (0.815 to 0.952)	0.844 (0.818 to 0.865)
SVM	0.944 (0.931 to 0.956)	0.921 (0.881 to 0.956)	0.854 (0.851 to 0.856)	0.945 (0.933 to 0.956)	0.869 (0.802 to 0.931)	0.858 (0.855 to 0.860)
MLP	0.975 (0.967 to 0.979)	1.00 (0.963 to 1.000)	0.922 (0.896 to 0.934)	0.867 (0.828 to 0.905)	0.500 (0.366 to 0.655)	0.925 (0.899 to 0.937)
RF	0.962 (0.953 to 0.970)	0.750 (0.684 to 0.815)	0.954 (0.950 to 0.958)	0.934 (0.920 to 0.946)	0.737 (0.647 to 0.824)	0.907 (0.902 to 0.912)
AB	1.000 (1.000 to 1.000)	1.000 (1.000 to 1.000)	1.000 (1.000 to 1.000)	0.499 (0.499 to 0.513)	0.000 (0.000 to 0.027)	0.999 (0.998 to 0.999)
LR	0.940 (0.926 to 0.953)	0.714 (0.650 to 0.774)	0.944 (0.943 to 0.946)	0.942 (0.928 to 0.954)	0.890 (0.835 to 0.944)	0.861 (0.859 to 0.863)

AB, boosted gradient trees; KNN, K-nearest neighbours; LR, logistic regression; MLP, neural network; RF, random forests; SVM, support vector machine.