Table 3.

Performance metrics.

Race	Feature selection model	Prediction method	TPR (sensitivities)	FPR	MCC	F1
AA	Boruta	Decision tree	0.7	0.166667	0.526789	0.682927
AA	Boruta	Random forest	0.909091	0.215686	0.559715	0.625
AA	Boruta	Logistic	0.733333	0.212766	0.471053	0.611111
AA	RFE	Random forest	0.833333	0.22	0.512001	0.606061
AA	RFE	SVM	0.8	0.25	0.427428	0.516129
AA	RFE	Decision tree	0.368421	0.176471	0.213946	0.4375
AA	Boruta	SVM	0.857143	0.272727	0.390796	0.428571
AA	RFE	Logistic	0.75	0.277778	0.334493	0.413793
NHW	Boruta	Random forest	0.8	0.285714	0.515515	0.761905
NHW	RFE	Logistic	0.736842	0.285714	0.450564	0.717949
NHW	Boruta	Logistic	0.647059	0.277778	0.370494	0.666667
NHW	Boruta	Decision tree	0.631579	0.4375	0.194079	0.631579
NHW	RFE	Decision tree	0.631579	0.4375	0.194079	0.631579
NHW	RFE	Random forest	0.833333	0.22	0.512001	0.606061
NHW	RFE	SVM	0.714286	0.428571	0.280976	0.606061
NHW	Boruta	SVM	0.611111	0.470588	0.140984	0.594595

Comparison of the performance of various combinations of feature selection and classification methods.