Table 2 Summary of metrics obtained for the cross-validation and test stages of the models, along with their respective optimal hyperparameters
Model | Optimal hyperparameters (random_state = 24) | Cross-validation results (cv = 10) | Test data results |
|---|---|---|---|
Logistic Regression | C: 0.3 | Accuracy = 0.63 | Accuracy = 0.75 |
max_iter: 50 | Precision = 0.62 | Precision = 0.75 | |
penalty: l1 | Recall = 0.62 | Recall = 0.75 | |
l1_ratio: 0.2 | F1-Score = 0.62 | F1-Score = 0.75 | |
solver: liblinear | |||
Gradient Boosting Classifier | n_estimators: 2000 | Accuracy = 0.59 | Accuracy = 0.80 |
learning_rate: 0.01 | Precision = 0.59 | Precision = 0.80 | |
criterion: friedman_mse | Recall = 0.59 | Recall = 0.80 | |
max_depth: 5 | F1-Score = 0.59 | F1-Score = 0.80 | |
loss: deviance | |||
K-Nearest Neighbors | n_neighbors: 1 | Accuracy = 0.75 | Accuracy = 0.80 |
weights: uniform | Precision = 0.75 | Precision = 0.80 | |
leaf_size: 1 | Recall = 0.75 | Recall = 0.80 | |
p: 5 | F1-Score = 0.75 | F1-Score = 0.80 | |
Support Vector Machine | kernel: rbf | Accuracy = 0.78 | Accuracy = 0.80 |
C: 0.9 | Precision = 0.78 | Precision = 0.80 | |
gamma: auto | Recall = 0.78 | Recall = 0.80 | |
F1-Score = 0.77 | F1-Score = 0.80 | ||
MLP Classifier | activation: relu | Accuracy = 0.73 | Accuracy = 0.75 |
alpha: 1.0 | Precision = 0.72 | Precision = 0.75 | |
hidden_layer_sizes: 1000 | Recall = 0.72 | Recall = 0.75 | |
learning_rate_init: 0.1 | F1-Score = 0.73 | F1-Score = 0.75 | |
max_iter: 50 | |||
solver: sgd | |||
Decision Tree | criterion: entropy | Accuracy = 0.66 | Accuracy = 0.85 |
max_depth: none | Precision = 0.66 | Precision = 0.85 | |
splitter: best | Recall = 0.66 | Recall = 0.85 | |
F1-Score = 0.66 | F1-Score = 0.85 | ||
Random Forest Classifier | max_depth: 10 | Accuracy = 0.70 | Accuracy = 0.85 |
n_estimators: 200 | Precision = 0.70 | Precision = 0.85 | |
min_samples_split: 2 | Recall = 0.70 | Recall = 0.85 | |
min_samples_leaf: 1 | F1-Score = 0.70 | F1-Score = 0.85 | |
criterion: gini | |||
max_features: auto |