Intelligent Fault Diagnosis of Bearing Based on Convolutional Neural Network and Bidirectional Long Short-Term Memory

Table 5

Comparison between other methods in the literature and our proposed model.

Classifier	Data preprocessing	CWRU test accuracy	Calculation time (seconds)
KNN [36]	HOCs and WT	91.23%
SVM [37]	WP	98.7% (4 classes)	12.914
RF [38]		98.04% (4 classes)
WT-CNN [26]	WT	99.25% (4 classes)
WT-GAN-CNN [26]	WT	100% (4 classes)
Compact 1D CNN [29]		93.2% (6 classes)
CNN-LSTM [39]		99.77% (6 classes)	419
CWT-CNN-RF [40]	CWT	99.73% (10 classes)	114.57
MSCNN-LSTM [32]		98.46% (10 classes)
CNN-BLSTM		100% (10 classes)	27.268