Compare Random Forests and Boosted Trees: Bias, Variance, Speed
Company: TikTok
Role: Data Scientist
Category: Machine Learning
Difficulty: medium
Interview Round: Technical Screen
##### Scenario
Product team wants to understand pros/cons of Random Forests versus Boosted Decision Trees and whether any feature scaling is required for these tree-based algorithms.
##### Question
Compare Random Forests and Gradient-Boosted Decision Trees (e.g., XGBoost) in terms of bias/variance, interpretability, training speed, and robustness to overfitting. Is feature standardization or normalization necessary for tree-based models? Explain why or why not.
##### Hints
Focus on ensemble construction, sequential vs. parallel learning, split criteria, and how trees handle monotonic transformations.
Quick Answer: This question evaluates understanding of ensemble machine learning methods—specifically differences between Random Forests and gradient-boosted decision trees—and the impact of feature preprocessing on tree-based models.