An Intelligent Fault Detection Model for Fault Detection in Photovoltaic Systems

Table 1

Parameters used in the MLP.

Parameters	Values
Algorithm	Stochastic gradient descent
Activation function	ReLU (input, hidden layers); Softmax (output layer)
Layer	3 hidden layers with 8 units each; 3 units at the output
Loss function	Categorical crossentropy
Optimizer	Adam
Data split	Train: 80%, test: 20%
Batch size	5 (with 200 epochs)
Tuning	-Fold crossvalidation, dropout