Home / ML Associate practice test / Model Development / Question

Databricks Machine Learning Associate · free practice question

Why do practitioners often prefer k-fold cross-validation over a single train-validation split when estimating model performance during development?

Exam-style question on Model Development from the free ML Associate practice test, with the answer and a full explanation below.

Updated November 10, 2026

Model Development

Why do practitioners often prefer k-fold cross-validation over a single train-validation split when estimating model performance during development?

  1. ABecause cross-validation avoids fitting the model repeatedly, which lowers the compute cost of each estimator it scores
  2. BBecause cross-validation averages over multiple folds, which gives a more stable estimate from limited data than one split
  3. CBecause cross-validation makes the final test score equal to the average training score across the folds it produced
  4. DBecause cross-validation removes the need for a held-out evaluation set once hyperparameter tuning is complete
  5. EBecause cross-validation is specific to classification tasks, which makes it more specialized than a simple split
Answer & explanation

Correct answer: B

WHY B: Cross-validation uses repeated train-test partitions, which generally yields a more stable estimate and uses limited data more effectively than a single split.

WHY NOT A: It fits the model once per fold, so it increases rather than decreases compute cost.

WHY NOT C: It gives no assurance that test and training scores will be equal.

WHY NOT D: A separate final evaluation set is still good practice after tuning.

WHY NOT E: Cross-validation also applies to regression and other settings.

Take the full ML Associate practice test free →

Would you pass today?

This question is one of 45 in the free, timed Databricks Machine Learning Associate practice test. Take it and get a section-by-section score.

Start the free test →