An engineer has four hyperparameters, each with many plausible values, and only enough budget to test a limited number of configurations. Which search method is usually more efficient than exhaustive grid search in this situation?
Answer & explanation
Correct answer: A
WHY A: Random search lets the user control the evaluation budget directly and is often more efficient than grid search in larger spaces. WHY NOT B: Grid search can waste trials on unimportant dimensions. WHY NOT C: Manual search is not systematically more efficient. WHY NOT D: Full enumeration is usually the least budget-friendly option. WHY NOT E: LeaveOneOut is a cross-validation scheme, not a parameter search method.