Explain Random Forest randomness and implications
Overview
This question evaluates a candidate's understanding of Random Forest ensemble mechanics—sources of randomness, their effects on bias and variance, hyperparameter impacts, evaluation choices (OOB vs cross-validation), feature-importance bias, and class-imbalance strategies—within the Machine Learning domain for binary classification.
Snapchat
Oct 13, 2025, 9:49 PM
Data Scientist
Onsite
Machine Learning
0
0
Random Forest — Rigor and Practical Choices
Context: You are building a binary classifier with a Random Forest. The dataset has 100,000 rows, 100 features, and a 5% positive rate. Answer the following:
- Sources of Randomness
- Enumerate the sources of randomness in Random Forests (e.g., bootstrap sampling, feature subsampling at each split, random tie-breaking, randomized split points). For each, explain its typical effect on bias and variance.
- Hyperparameters for the Given Dataset
- Propose reasonable values for n_estimators, max_depth, and max_features for the dataset above.
- Justify how max_features controls correlation among trees and the bias–variance trade-off.
- OOB Error vs 5-Fold Cross-Validation
- Compare out-of-bag (OOB) error with 5-fold cross-validation (CV).
- When can they disagree and why?
- Feature Importance Bias
- Explain why impurity-based importances are biased toward continuous or high-cardinality features.
- Propose a corrected approach (e.g., permutation importance with stratified shuffles and repeated runs) and justify your design.
- Class Imbalance Strategies
- Outline strategies for class imbalance (e.g., class_weight, threshold moving, balanced subsampling).
- Discuss consequences for probability calibration and decision thresholds.
Solution
Login required
Show