Table 5. Out-of-sample validation after... | Emerald Publishing

Table 5.

Out-of-sample validation after SMOTE application

Method	Accuracy	Specificity	Precision	Recall	F1-score	AUC
Panel A – SMOTE oversampling: financial ratios as input variables (15 features) – Out of sample validation
RF	0.85	0.87	0.78	0.82	0.80	0.92
GBT	0.80	0.77	0.68	0.87	0.76	0.91
DT	0.81	0.83	0.60	0.76	0.67	0.89
LR	0.50	0.25	0.41	0.95	0.58	0.85
SVM	0.45	0.16	0.39	0.96	0.56	N.A.
Panel B – SMOTE oversampling raw data of financial ratios as input variables (45 features) – Out of sample Validation
RF	0.91	0.89	0.82	0.94	0.88	0.97
GBT	0.87	0.84	0.75	0.93	0.83	0.96
DT	0.86	0.84	0.75	0.88	0.81	0.94
LR	0.82	0.82	0.71	0.81	0.76	0.87
SVM	0.79	0.80	0.67	0.77	0.72	N.A.
Panel C – SMOTE oversampling: raw data as input variables (215 features) – Out of sample Validation
RF	0.97	0.98	0.86	0.94	0.90	0.99
GBT	0.97	0.97	0.73	0.92	0.82	0.99
DT	0.96	0.95	0.98	0.96	0.97	0.99
LR	0.94	0.95	0.70	0.90	0.79	0.96
SVM	0.94	0.95	0.69	0.87	0.77	N.A.

Method	Accuracy	Specificity	Precision	Recall	F1-score	AUC
Panel A – SMOTE oversampling: financial ratios as input variables (15 features) – Out of sample validation
RF	0.85	0.87	0.78	0.82	0.80	0.92
GBT	0.80	0.77	0.68	0.87	0.76	0.91
DT	0.81	0.83	0.60	0.76	0.67	0.89
LR	0.50	0.25	0.41	0.95	0.58	0.85
SVM	0.45	0.16	0.39	0.96	0.56	N.A.
Panel B – SMOTE oversampling raw data of financial ratios as input variables (45 features) – Out of sample Validation
RF	0.91	0.89	0.82	0.94	0.88	0.97
GBT	0.87	0.84	0.75	0.93	0.83	0.96
DT	0.86	0.84	0.75	0.88	0.81	0.94
LR	0.82	0.82	0.71	0.81	0.76	0.87
SVM	0.79	0.80	0.67	0.77	0.72	N.A.
Panel C – SMOTE oversampling: raw data as input variables (215 features) – Out of sample Validation
RF	0.97	0.98	0.86	0.94	0.90	0.99
GBT	0.97	0.97	0.73	0.92	0.82	0.99
DT	0.96	0.95	0.98	0.96	0.97	0.99
LR	0.94	0.95	0.70	0.90	0.79	0.96
SVM	0.94	0.95	0.69	0.87	0.77	N.A.