Question 22
Domain 3: Train and evaluate modelsYou are creating a model to predict the price of a student’s artwork based on the student’s length of education, degree type, and art form. You start by building a linear regression model. Which evaluation metrics are appropriate for assessing this regression model? Does the proposed solution meet the goal?
Correct answer: A
Explanation
A linear regression model predicts a continuous value, so it should be evaluated with regression metrics such as "mean squared error," "root mean squared error," "mean absolute error," or "coefficient of determination (R-squared)." Classification metrics like "accuracy, precision, recall, F1 score, or AUC" measure class labels, so the proposed solution does not meet the goal.
Why each option is right or wrong
A. No. For a linear regression model, you should use regression metrics such as mean squared error, root mean squared error, mean absolute error, or coefficient of determination (R-squared), not classification metrics like accuracy, precision, recall, F1 score, or AUC.
Linear regression is used for a continuous target, so the proper evaluation criteria are regression losses and fit measures such as mean squared error, root mean squared error, mean absolute error, and R-squared. Classification metrics like accuracy, precision, recall, F1, and AUC are defined for discrete class labels, so using them here would not assess how close the predicted artwork prices are to the actual prices.
B. Yes. The listed metrics are all appropriate for evaluating a linear regression model because they measure how well the model predicts continuous values.