A student is asked to demonstrate underfitting in a hands-on exercise using a dataset with a clear nonlinear relationship. Which approach will most reliably produce underfitting?

Train a degree-10 polynomial with dropout regularization.

Train a linear regression model on the raw features without any polynomial expansion.

Train a random forest with 500 trees on the raw features.

Train a deep neural network with a large batch size and many epochs.

A team is designing a hands-on exercise to demonstrate overfitting. They have a small dataset of 100 samples with 50 features. Which of the following choices are likely to produce overfitting? Select all that apply.

Training a logistic regression model with only the top 3 features.

Training a decision tree with unlimited depth and no minimum samples per leaf.

Training a k-nearest neighbors model with k=1.

Training a linear regression model with strong L2 regularization.

Training a high-degree polynomial regression (degree 20) on the raw features.

Which technique is commonly used to mitigate underfitting?

Reducing the number of features.

Increasing the model's complexity, such as using a higher-degree polynomial.

Decreasing the number of training epochs.

Adding more regularization to the model.

A logistic regression model for customer churn achieves 62% accuracy on both training and test sets. The business requires at least 80% accuracy. Which approach is most likely to close this gap?

Apply stronger L2 regularization to the existing features.

Add interaction terms and polynomial features to the model.

Collect more training data and retrain the same model.

Use a simpler model such as a single decision stump.

A model's training accuracy is 55% and its test accuracy is 54% on a binary classification problem. What does this suggest?

The model has a high variance problem.

Certified Tester AI Testing

Domain 3Objective 9

Hands-On Exercise: Demonstrate Overfitting and Underfitting CT-AI Practice Questions (Page 2)

Part of the Domain 3: Machine Learning (ML) - Overview domain, which makes up ~15% of our current practice bank. ISTQB does not publish an official question count, but from its 60-minute exam (~25–40 total, ~4–6 in this domain), expect 1–1 from this objective — we provide 26 practice questions to prepare you well beyond it. (estimate)

26questions here

6free pages

11concepts

Questions 6–10

6application · medium
A student is asked to demonstrate underfitting in a hands-on exercise using a dataset with a clear nonlinear relationship. Which approach will most reliably produce underfitting?
Select an answer first
7expert · hard · select all that apply
A team is designing a hands-on exercise to demonstrate overfitting. They have a small dataset of 100 samples with 50 features. Which of the following choices are likely to produce overfitting? Select all that apply.
Select an answer first
8foundation · easy
Which technique is commonly used to mitigate underfitting?
Select an answer first
9application · medium
A logistic regression model for customer churn achieves 62% accuracy on both training and test sets. The business requires at least 80% accuracy. Which approach is most likely to close this gap?
Select an answer first
10foundation · easy
A model's training accuracy is 55% and its test accuracy is 54% on a binary classification problem. What does this suggest?
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.