Application of Feedforward Neural Network and SPT Results in the Estimation of Seismic Soil Liquefaction Triggering

Table 3

Hyperparameters space of machine learning (ML) models.

Hyperparameter	Range
Number of neurons	2 ÷ 100
Solver^(^)	1, 2, 3
Activation function^(^)	1, 2, 3
Max iteration	1000, 2000, 3000, 4000
Learning rate	0.001, 0.01, 0.05, 0.09, 0.1
Cost function	Accuracy
Data used	Training set/10-fold CV

⁽⁾: 1: quasi-Newton method; 2: stochastic gradient descent; 3: Adam. ⁽⁾: 1: logistic; 2: tanh; 3: ReLU.