A linear regression model has a high R-squared but the residuals show a clear pattern when plotted against fitted values. What is the most appropriate conclusion?

The model is definitely correct because R-squared is high

The model may be misspecified despite high R-squared

The residuals are normally distributed

The model has no multicollinearity

A data scientist is fitting a ridge regression model. After standardizing the features, they notice that the coefficients for two highly correlated features are both around 0.5. They want to reduce the model's variance further without losing the predictive contribution of either feature. What should they do?

Remove one of the correlated features.

Decrease the regularization strength alpha.

Increase the regularization strength alpha.

Which of the following is a key assumption of ordinary least squares (OLS) linear regression?

The independent variables must be normally distributed

The dependent variable must be binary

The residuals are normally distributed with a mean of zero

There must be no correlation between the independent variables and the residuals

A team is building a regression model with many features, including some that are highly correlated. They want to use elastic net because it handles both feature selection and multicollinearity. What is the key advantage of elastic net over lasso in this scenario?

Elastic net is faster to train than lasso.

Elastic net always produces a more interpretable model.

Elastic net can keep groups of correlated features together, while lasso tends to select only one from a correlated group.

Elastic net does not require feature scaling.

What is the primary purpose of adding an L2 penalty (ridge regularization) to a linear regression model?

To force some coefficients to be exactly zero, performing feature selection

To shrink the coefficients toward zero, reducing model complexity and overfitting

To eliminate the need for feature scaling in the model

To increase the bias of the model to improve interpretability of all coefficients

GIAC Machine Learning Engineer

Domain 2Objective 2

Regressions GMLE Practice Questions (Page 3)

Part of the Machine Learning Algorithms domain, which makes up ~29% of our current practice bank. GIAC (SANS) does not publish an official question count, but from its 180-minute exam (~70–120 total, ~20–35 in this domain), expect 7–12 from this objective — we provide 36 practice questions to prepare you well beyond it. (estimate)

36questions here

8free pages

5concepts

Questions 11–15

11expert · hard
A linear regression model has a high R-squared but the residuals show a clear pattern when plotted against fitted values. What is the most appropriate conclusion?
Select an answer first
12expert · hard
A data scientist is fitting a ridge regression model. After standardizing the features, they notice that the coefficients for two highly correlated features are both around 0.5. They want to reduce the model's variance further without losing the predictive contribution of either feature. What should they do?
Select an answer first
13foundation · easy
Which of the following is a key assumption of ordinary least squares (OLS) linear regression?
Select an answer first
14expert · hard
A team is building a regression model with many features, including some that are highly correlated. They want to use elastic net because it handles both feature selection and multicollinearity. What is the key advantage of elastic net over lasso in this scenario?
Select an answer first
15foundation · easy
What is the primary purpose of adding an L2 penalty (ridge regularization) to a linear regression model?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to GMLE

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 GIAC (SANS). “GMLE” is a trademark of its owner, used for identification only.