Table 3.

Measures of confusion matrix.

Confusion matrix elements	Original dataset (OR-Sample)	Under-sample dataset (U-sample)	Over-sample dataset (O-sample)	Over-under sample dataset (OU-sample)	ROSE sample dataset (R-sample)
Accuracy	0.9035	0.5164	0.9729	0.9442	0.6618
95% CI	0.8949, 0.9115	0.4724, 0.5603	0.9696, 0.9758	0.9378, 0.9502	0.6485, 0.6749
No information rate	0.9488	0.5551	0.5838	0.5334	0.5107
p-value [Acc>NIR]	1.000	0.96495	<2.2e-16	<2.2e-16	<2.2e-16
Specificity	0.10156	0.5017	0.9982	0.9825	0.6669
Balanced accuracy	0.52413	0.5183	0.9765	0.9466	0.6617
Precision	0.0932	0.5017	0.9548	0.9107	0.6565
Recall	0.10	0.5737	0.9987	0.9835	0.6538
F1	0.0972	0.5183	0.9762	0.9457	0.6551

Source: Authors' own computation.