Table 4.

Gender- and age-stratified models. Three separate data sets were constructed: a male data set with male participants, a female data set with female participants, and age-trimmed data set by excluding the participants younger than 50 years. For each of these data sets, a separate model was constructed, and its performance is reported below (N=726).

Algorithm	Male (n=415)		Female (n=477)			Age-trimmed (male, n=366 and female, n=426)
	AUCa	Accuracy	AUC	Accuracy	AUC		Accuracy
SVMb	0.795	0.717	0.659	0.763	0.755		0.723
Random Forest	0.758	0.702	0.699	0.788	0.739		0.713
LightGBM	0.725	0.665	0.717	0.768	0.749		0.712
XGBoost	0.762	0.717	0.682	0.771	0.742		0.704

^aAUC: area under the curve.

^bSVM: support vector machine.