Prediction of the Control Effect of Fractured Leakage in Unconventional Reservoirs Using Machine Learning Method

Table 3

Details of the developed artificial neural network architecture.

Feature	Value/model
Number of samples	450
Training algorithm	Levenberg-Marquardt algorithm
Hidden layer size
Tolerance
Maximum iteration	2000
Learning rate	0.07
Initial learning rate	0.01
Activation	Logistic sigmoid functions
Shuffle	TURE