Model Training and Prediction
Random Forest Learner:
- Target: Churn
- 100 trees (default: balances accuracy & speed)
- Info Gain Ratio
- Trained on 80% training set (800 rows)
Random Forest Predictor:
- Applied to 20% test set (200 rows)
- Outputs: predicted label + P(Churn=Yes)
- Probability columns used for ROC analysis
Why Random Forest?
Handles non-linearity, robust to outliers,
and provides class probabilities for threshold tuning.
