Intelligent Diagnosis of Rolling Bearing Fault Based on Improved Convolutional Neural Network and LightGBM

Table 1

Hyperparameter setting of the improved convolution neural network.

Names of the hyper parameters	Parameter value
Input random deactivation rate	0.2
Loss function	Categorical_crossentropy
Optimizer	Adam
Initial learning rate	0.001
Attenuation coefficient	0.5
Attenuation interval	5
Minimum learning rate	0.00001
Number of batches	32
Iterations	20