A ResNet-LSTM Based Credit Scoring Approach for Imbalanced Data

Table 6

Classification method results of model.

Method	Bank				UCI
	Recall	F1	AUC	KS	Recall	F1	AUC	KS
RNN-RF	82.07	78.44	87.89	70.64	25.77	35.60	60.24	20.28
Adaboost	98.73	78.89	95.25	87.25	31.89	43.55	63.92	27.66
SMOTEBoost	81.33	73.69	87.52	72.39	39.08	46.88	65.69	29.16
CUSBoost	74.05	70.22	84.09	64.50	36.33	46.11	65.21	27.59
RUSBoost	96.68	78.84	94.47	88.91	61.43	51.72	70.12	40.18
CNN	80.70	73.38	87.22	88.54	39.39	47.13	65.83	40.11
LSTM	95.09	79.03	93.91	89.46	28.01	38.10	61.36	34.05
Propose	98.41	82.52	96.93	89.54	57.34	67.91	75.40	53.29