Safe Operation Management of Urban Smart Grid Based on Deep Learning

Table 3

Fifty-fold cross-validation dataset.

Parameters	Values	Parameters	Values
Optimizer	SGD	Learning rate	0.01
Batch size	32	Number of training rounds	10
Number of neural network layers	5	Input units	86
Hidden layer units	128, 64, 32	Dropout	0.1
Activation functions	Tanh	Pretraining optimizer	SGD