Explain K-Fold Cross-Validation and Its Trade-Offs
Company: Amazon
Role: Data Scientist
Category: Machine Learning
Difficulty: medium
Interview Round: Technical Screen
##### Scenario
Phone interview focusing on breadth and depth of machine-learning knowledge.
##### Question
Explain the k-fold cross-validation procedure. What are the trade-offs in choosing the number of folds, and when might cross-validation give misleading performance estimates?
##### Hints
Address bias-variance, data leakage, temporal data, and class imbalance effects.
Quick Answer: This question evaluates a candidate's understanding of k-fold cross-validation and related competencies in model evaluation, including reasoning about bias–variance trade-offs, computational cost, and pitfalls such as data leakage, temporal/ordered data, and class imbalance.
