Table 2 Detailed learning parameters and cross-validation setup for all classification models.
Classifier | Architecture / parameters | Criterion / Kernel | Regularization / Ensemble | Activation function | Optimizer & loss function | Cross-validation |
|---|---|---|---|---|---|---|
SVM | C = 1.0, gamma = ‘scale’, probability = True | RBF kernel | Soft-margin regularization (C = 1.0) | – | – | 5-fold |
DT | random_state = 42 | Gini impurity | No regularization | – | – | 5-fold |
RF | n_estimators = 100, criterion = ‘gini’, random_state = 42 | Gini impurity | Ensemble of 100 decision trees | – | – | 5-fold |
DNN | 4 layers: Input (8) → Dense (64) → Dropout (0.3) → Dense (32) → Dropout (0.2) → Dense (16) → Dense (1) | – | Dropout (0.3 and 0.2), EarlyStopping (patience = 10) | ReLU (hidden layers), Sigmoid (output) | Adam optimizer (learning rate = 0.0001), Binary Cross-Entropy Loss | 5-fold |
