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

Table 1

Features expression.

Label	Expression	Label	Expression
Peak factor		Form factor
Margin factor		Power spectrum gravity index
Pulse factor		Power spectrum variance
Kurtosis		Mean square frequency
Skewness		Information entropy

The details of the expressions in the table are as follows: is frequency at th sample; is power spectrum at th sample; is density.