Modified Kernel Marginal Fisher Analysis for Feature Extraction and Its Application to Bearing Fault Diagnosis

Table 2

The recognition rates (%) of KNN classifier based on six feature extraction approaches.

Fault type	⁢Methods
	KMFA; ()	KPCA; ()	KFDA; ()	SKLPP; ()	RKMFA; ()	MKMFA; ()
Inner race fault	98.89	80.00	100	100	100	100
Ball fault	6.67	6.67	97.78	100	100	100
Outer race fault	100	100	100	100	100	100
Normal	100	13.34	15.56	13.34	88.89	100