Rolling Bearing Fault Diagnosis Based on Domain Adaptation and Preferred Feature Selection under Variable Working Conditions

Table 1

A list of acronyms.

Acronym	Full form
TCAPLMS	Transfer component analysis with preserving local manifold structure
PSFFC	Preferred feature selection by fault sensitivity and feature correlation
REBs	Rolling element bearings
MODWPT	Maximum overlap discrete wavelet packet transform
SVD	Singular value decomposition
DTCWPT	Dual-tree complex wavelet packet transform
VMD	Variational mode decomposition
IMF	Intrinsic mode function
FAWT	Flexible analytical wavelet transform
TF	Time-frequency
STFT	Short-time Fourier transform
CNN	Convolutional neural network
EWT	Empirical wavelet transform
K	Kurtosis
PV	Peak value
V	Variance
RMS	Root mean square
LFDA	Local Fisher discriminant analysis
HES	HHT envelope spectrum
OFS	Original feature set
WPD	Wavelet packet denoising
CWRU	Case Western Reserve University
HES	Hilbert envelope spectra
RF	Random forest
SNR	Signal-to-noise ratio
EMD	Empirical mode decomposition
STD	Standard deviation
WSVM	Weighted support vector machine
CDET	Compensation distance evaluation technique
LPP	Locality preserving projections
TCA	Transfer component analysis
HKL	High-order Kullback–Leibler
MSDCTL	Multistage deep convolutional transfer learning
DWT	Discrete wavelet transform
MODWT	Maximal overlap discrete wavelet transform
LDA	Linear discriminant analysis
Sw	Skewness
SMD	Sum of within-class mean deviations
MMD	Maximum mean discrepancy
MD	Mean deviations
ARI	Adjusted rand index
PCC	Pearson correlation coefficient
FPSD	Feature priority selection degree
FSD	Fault sensitivity degree
RFS	Raw feature set