Credit Rating Using Type-2 Fuzzy Neural Networks

Table 3

Comparative results of different credit rating models.

Methods	Number of hidden neurons	Epochs	RMSE		Recognition rate
			Train	Test	Training	Testing
SVM	—				86	88.54
Neural networks	6	5000	0.111803	0.412263	98.2500	83.003953
	16	5000	0.110000	0.417029	98.7500	82.608696
	32	5000	0.122474	0.392620	98.5000	84.584980
Fuzzy NN	6	2000	0.226731	0.444709	95.5000	84.189723
	16	2000	0.231411	0.250698	93.5000	92.50000
	32	2000	0.219856	0.229424	95.7500	93.25000
Type-2 TSK	16	1000	0.206510	0.229653	96.2500	93.313834