🔁 Five-Fold Cross-Validation

Five-fold cross-validation (5-fold CV) is a statistical method used to evaluate the performance of a machine learning model. It helps ensure the model generalizes well to unseen data.

* The dataset is randomly split into 5 equal parts (folds). * The model is trained on 4 folds and tested on the remaining 1 fold. * This process repeats 5 times, each time using a different fold as the test set. * Final performance is the average of the 5 results.

* Reduces overfitting risk * Provides a more reliable estimate of model performance than a single train/test split * Especially useful when data is limited

1. Total samples = 100 patients with MRI radiomics
2. Split into 5 folds of 20 patients each:
   • Iteration 1: Train on folds 1–4, test on fold 5
   • Iteration 2: Train on folds 1,2,3,5, test on fold 4
   • …
   • Iteration 5: Train on folds 2–5, test on fold 1
3. Compute metrics (e.g., accuracy, AUC) at each step
4. Final result: mean of all 5 metrics

• Uses all data for both training and testing
• Provides variance estimate of model performance
• Good balance between bias and variance

• Use stratified 5-fold CV when class labels are imbalanced
• Avoid data leakage: normalize or extract features within each fold

from sklearn.model_selection import cross_val_score
from xgboost import XGBClassifier
 
model = XGBClassifier()
scores = cross_val_score(model, X, y, cv=5, scoring='accuracy')
 
print("5-fold CV Accuracy:", scores.mean())