Table 2 Optimized hyperparameters of the machine learning models for predicting the CS of SFRC CS.
Machine learning model | Optimized values/types of key hyperparameters |
|---|---|
KNN | n_neighbors = 5 |
SVR | kernel = RBF C = 6 gamma = 0.2 epsilon = 0.1 |
GPR | kernel = ConstantKernel(50)*RBF(1) + RBF(1) alpha = 1 random_state = 1 |
RFR | n_estimators = 200 max_depth = 70 max_features = 9 min_samples_split = 2 min_samples_leaf = 1 bootstrap = True |
XGBR | colsample_bytree = 0.7 n_estimators = 200 learning_rate = 0.02 max_depth = 8 |
ANN | hidden layers = 2 number of neurons = 1 st hidden layer: 64, 2nd hidden layer: 32 activation = ReLU optimizer = Adam learning rate = 0.001 batch size = 32 epochs = 100 loss function = MSE |
