Feature Description

Implement feature importance calculation for both RandomForestClassifier and RandomForestRegressor. Feature importance helps users understand which features contribute most to predictions.

Implementation Requirements

Core Methods:

• feature_importances(&self) -> Option<Vec<f32>> - Returns importance scores for each feature
• Based on Gini impurity decrease (classification) or variance reduction (regression)
• Aggregate importances across all trees in the forest

Algorithm:

1. For each tree, track impurity decrease at each split
2. Accumulate importance by feature index
3. Normalize: importance_i = total_decrease_i / sum(all_decreases)
4. Average across all trees

Testing:

• Comprehensive tests for both classifier and regressor
• Verify importances sum to 1.0
• Test with known feature relationships
• Verify reproducibility with random_state

Examples:

• Update examples/random_forest_iris.rs to show feature importance
• Update examples/random_forest_regression.rs to show feature importance
• Add visualization of top features

Documentation:

• Add theory section to book/src/ml-fundamentals/ensemble-methods.md
• Explain Gini importance vs permutation importance
• When to use feature importance for model interpretation

Acceptance Criteria

• feature_importances() method for RandomForestClassifier
• feature_importances() method for RandomForestRegressor
• Importance scores sum to 1.0
• 8+ comprehensive tests
• Updated examples showing feature importance
• Documentation with theory and usage
• Zero clippy warnings
• All tests pass

References

• Breiman (2001): Random Forests - Section on variable importance
• sklearn RandomForestClassifier.feature_importances_