Question 8
Content Domain 3: ModelingA Machine Learning Specialist trained a regression model, but the first iteration needs optimizing. The Specialist needs to understand whether the model is more frequently overestimating or underestimating the target. What option can the Specialist use to determine whether it is overestimating or underestimating the target value?
Correct answer: B
Explanation
Residual plots show the difference between predicted and actual values, so they reveal whether errors are mostly positive or negative. In regression, positive residuals mean the model is underestimating the target, while negative residuals mean it is overestimating it. This makes residual plots useful for spotting bias in the first iteration.
Why each option is right or wrong
A. Root Mean Square Error (RMSE).
RMSE summarizes overall error size but does not show whether errors are mostly positive or negative.
B. Residual plots.
Residual plots are the standard diagnostic for regression error direction because they graph the residual, defined as actual value minus predicted value, for each observation. Under the usual convention, residuals above 0 indicate the model predicted too low and residuals below 0 indicate it predicted too high, so the plot immediately shows whether the errors are skewed positive or negative across the sample.
C. Area under the curve.
Area under the curve is mainly a classification metric, not a regression bias diagnostic.
D. Confusion matrix.
Confusion matrices evaluate classification outcomes, not continuous prediction overestimation or underestimation.