Threshold optimization and random undersampling for imbalanced credit card data

Journal of Big Data

Table 33 Results for the XGBoost depth 1 classifier

Technique	Threshold	TPR	FPR	FNR	TNR	F-meas	G-mean	MCC	Precision
F-meas	0.0241	0.9170	0.0384	0.0830	0.9616	0.0767	0.9390	0.1870	0.0400
F-meas NC	0.3270	0.8551	0.0015	0.1449	0.9985	0.6273	0.9238	0.6510	0.4996
G-mean	0.0212	0.9219	0.0451	0.0781	0.9549	0.0666	0.9382	0.1736	0.0346
G-mean NC	0.0335	0.9057	0.0280	0.0943	0.9720	0.1095	0.9381	0.2237	0.0587
MCC	0.0211	0.9219	0.0452	0.0781	0.9548	0.0665	0.9381	0.1733	0.0345
MCC NC	0.0365	0.9042	0.0267	0.0958	0.9733	0.1199	0.9379	0.2334	0.0653
Precision	0.0241	0.9170	0.0384	0.0830	0.9616	0.0767	0.9390	0.1870	0.0400
Precision NC	0.9284	0.6142	0.0002	0.3858	0.9998	0.7054	0.7760	0.7261	0.8842
C	0.0357	0.9046	0.0234	0.0954	0.9766	0.1177	0.9398	0.2350	0.0630
D	0.5000	0.8429	0.0008	0.1571	0.9992	0.7294	0.9175	0.7362	0.6452