Which of the following correctly lists the typical stages of a training pipeline in the order they are executed?

Data preprocessing → feature engineering → model training → evaluation

Feature engineering → data preprocessing → evaluation → model training

Model training → data preprocessing → feature engineering → evaluation

Evaluation → model training → feature engineering → data preprocessing

A team is working with a highly imbalanced classification dataset where the positive class is only 2% of the data. They want to use cross-validation to evaluate a model. What is the most appropriate cross-validation strategy?

StratifiedKFold with shuffle=True.

Standard KFold with shuffle=True and random_state=42.

A single train-test split with a 90/10 ratio.

A team is comparing two classification models. They have a small dataset of 2,000 rows and want a performance estimate that is less sensitive to which particular rows end up in the validation set. They also want to see per-fold scores to understand variability. Which approach should they use?

Train on all 2,000 rows and evaluate on the same data to maximize sample size.

Use cross_validate with 5-fold CV and examine the test_score array.

Perform a single 80/20 train-test split and compute accuracy on the test set.

Use cross_val_score but only look at the mean, ignoring individual fold results.

A data scientist wants to evaluate a logistic regression model using cross-validation. They also want to capture multiple metrics (accuracy, precision, recall) and the training time for each fold. Which scikit-learn function should they use?

cross_validate with scoring=['accuracy', 'precision', 'recall'] and return_train_score=True.

GridSearchCV with scoring='accuracy'.

train_test_split followed by manual evaluation.

cross_val_score with scoring='accuracy'.

Which of the following is a complexity introduced by k-fold cross-validation that is not present in a single train-validation split?

The need to choose the number of folds (k)

The need to encode categorical variables

The need to shuffle the data before splitting

DatabricksCertified Machine Learning Associate

Domain 3Objective 3

Training Pipelines and Cross-Validation MACHINE-LEARNING-ASSOCIATE Practice Questions (Page 4)

Part of the Section 3: Model Development domain, which makes up ~23% of our current practice bank. Databricks does not publish an official question count, but from its 90-minute exam (~35–60 total, ~8–14 in this domain), expect 2–4 from this objective — we provide 29 practice questions to prepare you well beyond it. (estimate)

29questions here

6free pages

9concepts

Questions 16–20

16foundation · easy
Which of the following correctly lists the typical stages of a training pipeline in the order they are executed?
Select an answer first
17expert · hard
A team is working with a highly imbalanced classification dataset where the positive class is only 2% of the data. They want to use cross-validation to evaluate a model. What is the most appropriate cross-validation strategy?
Select an answer first
18application · medium
A team is comparing two classification models. They have a small dataset of 2,000 rows and want a performance estimate that is less sensitive to which particular rows end up in the validation set. They also want to see per-fold scores to understand variability. Which approach should they use?
Select an answer first
19application · medium
A data scientist wants to evaluate a logistic regression model using cross-validation. They also want to capture multiple metrics (accuracy, precision, recall) and the training time for each fold. Which scikit-learn function should they use?
Select an answer first
20foundation · easy
Which of the following is a complexity introduced by k-fold cross-validation that is not present in a single train-validation split?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to MACHINE-LEARNING-ASSOCIATE

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by Databricks. “MACHINE-LEARNING-ASSOCIATE” is a trademark of its owner, used for identification only.