The Hybrid KICA-GDA-LSSVM Method Research on Rolling Bearing Fault Feature Extraction and Classification

Table 6

The performances of 4 commonly different classifiers using proposed new sensitive features set.

Method	LSSVM	KICA-LSSVM	LDA-LSSVM	GDA-LSSVM	KICA-GDA-LSSVM
Accuracy	69.32%	71.16%	71.731%	81.37%	92.17%