An ML engineer is building a binary classifier for a rare disease with only 2,000 positive samples and 98,000 negative samples. They need to tune hyperparameters and get a reliable final performance estimate. Which data splitting strategy is most appropriate?

Stratified split into 70% training, 15% validation, and 15% test sets.

Random split into 60% training, 20% validation, and 20% test sets.

Stratified split into 90% training and 10% test sets, using the test set for both tuning and evaluation.

Random split into 80% training and 20% test sets only.

A team is preparing a dataset for a medical diagnosis model. The 'blood_pressure' column has some values recorded as '120/80', '120 over 80', and '120-80'. The 'cholesterol' column has 15% missing values. What is the most appropriate cleaning step?

Parse 'blood_pressure' into systolic and diastolic numeric columns, and impute missing cholesterol with the median.

Convert all 'blood_pressure' values to a single string format and impute cholesterol with 0.

Replace all 'blood_pressure' values with the most common format and drop rows with missing cholesterol.

Drop the 'blood_pressure' column and impute missing cholesterol with the mean.

What is the purpose of scaling or normalization in data transformation?

To reduce the number of features in the dataset

To convert categorical variables into numerical ones

To bring numerical features to a similar range or scale

To handle missing values in numerical columns

A team is preparing a dataset for a churn prediction model. The 'customer_tenure' column has 20% missing values. The team is considering two approaches: impute with the median or build a separate model to predict the missing values. The dataset has 500,000 rows. Which approach is more appropriate given the constraints?

Build a separate model to predict tenure because it will always produce more accurate imputations.

Drop all rows with missing tenure.

Impute with the median because it is simple, fast, and the missingness is likely random.

Impute with 0 because missing tenure means the customer is new.

What is the primary purpose of data preparation in an ML pipeline?

To increase the size of the dataset by generating synthetic samples

To automatically select the best ML algorithm for the task

To transform raw data into a clean, consistent, and usable format for model training

To visualize data for business stakeholders

Certified Tester AI Testing

Domain 4Objective 2

Hands-On Exercise: Data Preparation for ML CT-AI Practice Questions (Page 4)

Part of the Domain 4: ML - Data domain, which makes up ~12% of our current practice bank. ISTQB does not publish an official question count, but from its 60-minute exam (~25–40 total, ~3–5 in this domain), expect 1–1 from this objective — we provide 25 practice questions to prepare you well beyond it. (estimate)

25questions here

5free pages

6concepts

Questions 16–20

16application · medium
An ML engineer is building a binary classifier for a rare disease with only 2,000 positive samples and 98,000 negative samples. They need to tune hyperparameters and get a reliable final performance estimate. Which data splitting strategy is most appropriate?
Select an answer first
17application · medium
A team is preparing a dataset for a medical diagnosis model. The 'blood_pressure' column has some values recorded as '120/80', '120 over 80', and '120-80'. The 'cholesterol' column has 15% missing values. What is the most appropriate cleaning step?
Select an answer first
18foundation · easy
What is the purpose of scaling or normalization in data transformation?
Select an answer first
19expert · hard
A team is preparing a dataset for a churn prediction model. The 'customer_tenure' column has 20% missing values. The team is considering two approaches: impute with the median or build a separate model to predict the missing values. The dataset has 500,000 rows. Which approach is more appropriate given the constraints?
Select an answer first
20foundation · easy
What is the primary purpose of data preparation in an ML pipeline?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to CT-AI

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 ISTQB. “CT-AI” is a trademark of its owner, used for identification only.