Credit Risk Evaluation with a Least Squares Fuzzy Support Vector Machines Classifier

Table 4

McNemar’s test for pairwise performance comparison of different models.

Model	LS-SVM	FSVM	SVM	ANN	LogR	LinR
LS-FSVM	0.0000 [1.0000]	0.0480 [0.8262]	15.488 [0.0001]	26.881 [0.0001]	59.274 [0.0001]	92.160 [0.0001]
LS-SVM		0.0120 [0.9131]	14.675 [0.0001]	25.840 [0.0001]	57.840 [0.0001]	90.482 [0.0001]
FSVM			13.133 [0.0003]	23.842 [0.0001]	55.048 [0.0001]	87.197 [0.0001]
SVM				1.5370 [0.2501]	15.584 [0.0001]	36.300 [0.0001]
ANN					7.0840 [0.0078]	22.781 [0.0001]
LogR						4.3630 [0.0367]