Skip to main content

View full-text article in PMC

. 2021 May 6;9(5):e22591. doi: 10.2196/22591

Table 3.

Performance of each model with all features input.

Model	Accuracy	AUCROC^a	Sensitivity	Specificity	Precision	F1
Random forests	0.914	0.986	0.877	0.955	0.955	0.914
Decision trees	0.792	0.796	0.712	0.881	0.867	0.782
kNN^b	0.743	0.779	0.712	0.776	0.776	0.743
LDA^c	0.829	0.882	0.781	0.881	0.877	0.826
AdaBoost^d	0.886	0.969	0.822	0.955	0.952	0.882
DNN^e	0.921	0.964	0.904	0.940	0.943	0.923

^aAUROC: area under the receiver operating characteristic curve.

^bkNN: k-nearest neighbor.

^cLDA: linear discriminant analysis.

^dAdaBoost: adaptive boosting.

^eDNN: deep neural network.