Random Forest averages many decision trees built on bootstrap samples to produce a strong, low‑variance model.
Key Parameters
- n_estimators, max_depth, max_features
- min_samples_split/leaf for regularization
- class_weight for imbalance
Validation
Use out‑of‑bag (OOB) estimates and a holdout or cross‑validation for final metrics. Keep preprocessing within folds.
Explainability
Permutation importances and partial dependence plots help communicate drivers; impurity importances provide a quick sanity check.
When to Use
As a strong baseline on tabular problems, or when you need performance with reasonable explainability and minimal tuning.