[Retracted] Prediction of the Least Principal Stresses Using Drilling Data: A Machine Learning Application

Table 2

Summary of the optimization process of the developed ANN-based model.

Parameter	Optimized parameters
	σmin model	σmax model
Number of hidden layers	Single hidden layer
Number of neurons in each layer	15	35
Split ratio	(Training/testing); 0.8/0.2
Training algorithms	Trainlm
Transfer function	Logsig	Tansig
Learning rate	0.12