Table 3 Comparison of diagnosis accuracy, R2, and MAE across different models and equipment types.

From: Bearing fault detection with lightweight feature extraction mechanism based on smoothed dilated convolution

Method

Accuracy (%)

R2

MAE

BPNN-UKF

88.34

0.821

0.156

TCN-BiLSTM

90.12

0.857

0.132

VMD-KELM

89.77

0.843

0.141

SDC-Shuffle-BiGRU

96.42

0.931

0.072

Back to article page