Table 1.

Model learning (cross validation results K = 10) of all the models tested in the different experiments.

	PCs used	F1-Score		Sensitivity (TPR)		Specificity (TNR)		Accuracy
		XGBOOST	SVM	XGBOOST	SVM	XGBOOST	SVM	XGBOOST	SVM
Experiment 1	n = 5	0.625 ± 0.041	0.595 ± 0.073	0.658 ± 0.093	0.780 ± 0.115	0.684 ± 0.062	0.580 ± 0.113	0.677 ± 0.043	0.626 ± 0.078
	n = 10	0.620 ± 0.043	0.639 ± 0.056	0.580 ± 0.084	0.580 ± 0.110	0.716 ± 0.066	0.739 ± 0.065	0.678 ± 0.043	0.695 ± 0.056
	n = 50	0.587 ± 0.045	0.558 ± 0.057	0.283 ± 0.098	0.238 ± 0.094	0.890 ± 0.055	0.891 ± 0.033	0.679 ± 0.065	0.703 ± 0.048
Experiment 2 (VTM1 results)	n = 5	0.749 ± 0.066	0.759 ± 0.086	0.538 ± 0.168	0.525 ± 0.150	0.946 ± 0.033	0.950 ± 0.051	0.852 ± 0.039	0.832 ± 0.070
	n = 10	0.793 ± 0.116	0.826 ± 0.088	0.646 ± 0.242	0.652 ± 0.176	0.939 ± 0.070	0.969 ± 0.031	0.884 ± 0.069	0.879 ± 0.065
	n = 50	0.754 ± 0.132	0.700 ± 0.160	0.548 ± 0.275	0.487 ± 0.319	0.961 ± 0.058	0.937 ± 0.071	0.876 ± 0.075	0.845 ± 0.080
Experiment 2 (VTM2 results)	n = 5	0.433 ± 0.080	0.418 ± 0.104	0.273 ± 0.081	0.285 ± 0.200	0.618 ± 0.119	0.588 ± 0.196	0.494 ± 0.082	0.491 ± 0.139
	n = 10	0.482 ± 0.106	0.515 ± 0.110	0.241 ± 0.162	0.381 ± 0.206	0.762 ± 0.138	0.688 ± 0.133	0.581 ± 0.111	0.594 ± 0.107
	n = 50	0.511 ± 0.120	0.476 ± 0.093	0.158 ± 0.172	0.135 ± 0.148	0.914 ± 0.063	0.925 ± 0.075	0.711 ± 0.083	0.654 ± 0.098
Experiment 3	n = 5	0.891 ± 0.109	0.979 ± 0.048	0.922 ± 0.127	0.974 ± 0.076	0.872 ± 0.157	0.988 ± 0.054	0.897 ± 0.109	0.980 ± 0.044
	n = 10	0.950 ± 0.050	0.947 ± 0.066	0.930 ± 0.070	0.987 ± 0.040	0.970 ± 0.090	0.918 ± 0.110	0.981 ± 0.052	0.950 ± 0.063
	n = 50	0.868 ± 0.109	0.968 ± 0.055	0.911 ± 0.106	0.989 ± 0.484	0.832 ± 0.214	0.941 ± 0.118	0.882 ± 0.093	0.972 ± 0.045
Robustness analysis	n = 5	0.647	0.964	0.956	1.000	0.377	0.923	0.687	0.964
	n = 10	0.609	0.945	0.922	0.964	0.346	0.923	0.654	0.945
	n = 50	0.690	0.902	0.956	1.000	0.446	0.795	0.719	0.902

Bold values represent the best model for each experiment using the selected metric F1-score.