An Integrated Method of Rolling Bearing Fault Diagnosis Based on Convolutional Neural Network Optimized by Sparrow Optimization Algorithm

Table 5

Optimization process of SAA from iteration 3 to 7.

Iteration	Batch size	Learning rate (keeping two significant digits)
3	160	0.48
	160	1.7
	160	0.0096
	160	0.0096
	4	0
4	100	0
	160	0
	160	0.0091
	160	0.0096
	4	0.78
5	4	0.27
	80	0.0073
	160	0.0096
	160	0.0099
	4	0.18
6	80	0.49
	160	0.23
	64	0.0066
	80	0.0073
	100	0.036
7	80	0.014
	80	0.0075
	40	0
	80	0.0070
	4	0