Skip to main content
. 2022 Nov 29;17(11):e0278217. doi: 10.1371/journal.pone.0278217

Table 4. Hyperparameter tuning of machine learning algorithms.

Algorithm Parameter Configuration Value
Logistic Regression {penalty, solver, C, max_iter} {‘l2’, ‘liblinear’, ‘1.0’, ‘100’}
Random Forest {criterion, n_estimators} {‘gini’, ‘100’}
Gradient Boosting {criterion, n_estimators, learning_rate} {‘friedman_mse’, ‘100’,’0.1’}
XGBoost {booster, gamma, n_estimators, learning_rate} {‘gbtree’, ‘1’, ‘100’,’0.1’}
Deep Neural Network {epoch, batch_size, activation, loss, network layer} {‘300’, ‘100’, ‘relu’, ‘binary_crossentropy’, [12-50-50-50-1]}
1D-CNN {epoch, batch_size, activation, loss, network layer} {‘300’, ‘100’, ‘relu’, ‘categorical_crossentropy, [25–32–32–64–64–64–64–2]}