Regression Accuracy Metrics (Part 2): Machine Learning Interview Prep 20

Shahidullah Kawsar
4 min readJan 12, 2024

Regression accuracy metrics are like report cards for how well our predictions match reality in a math problem. They tell us if our guesses are close to the actual answers or if we’re way off track. Regression accuracy metrics are essential tools used to assess the performance of regression models, which predict continuous outcomes. These metrics quantify the difference between predicted values and actual values, thereby evaluating the model’s accuracy. These metrics guide model refinement and selection, ensuring reliable predictions.

Photo: Twin Lakes, Colorado, USA. Credit: Tasnim and Kawsar

Let’s check your knowledge of Supervised Learning — Regression Accuracy Metrics. Here are 10 multiple-choice questions for you and there’s no time limit. Have fun!

1. Which metric would be best for the dataset containing many outliers?
(A) Mean Absolute Error (MAE)
(B) Mean Squared Error (MSE)
(C) R-squared
(D) Root Mean Squared Error (RMSE)

2. What is the main disadvantage of using R-squared in evaluating regression models?
(A) It can be negatively affected by outliers.
(B) It can misleadingly increase with the addition of more predictors, even if they are not relevant to the model, leading to overfitting.
(C) It measures absolute, not relative, error.
(D) It does not indicate the direction of the error.

3. RMSE is the sample standard deviation of the differences between predicted and observed values. In the context of regression analysis, what is the advantage of using RMSE over MSE?
(A) RMSE is less sensitive to outliers.
(B) RMSE is easier to interpret in the context of the original data.
(C) RMSE provides a more conservative estimate.
(D) RMSE is always lower than MSE.

4. What is the primary purpose of Mean Absolute Error (MAE) in regression analysis?
(A) To square the errors before summing them up.
(B) To measure the average magnitude of the errors.
(C) To amplify the effect of outliers in the dataset.
(D) To assess the categorical data accuracy.

5. What does an R-squared value of 1 indicate in a regression model?
(A) No correlation between variables.
(B) A perfect fit to the data.
(C) High bias in the model.
(D) Only a random relationship between variables.

6. Mean Absolute Percentage Error (MAPE) can become undefined or misleadingly high. When is MAPE not suitable?
(A) The dataset contains zero or near-zero actual values.
(B) The dataset has outliers.
(C) The model is a linear regression.
(D) The dependent variable is continuous.

7. When Adjusted R-squared is suitable?
(A) It penalizes the addition of irrelevant predictors.
(B) It works better with large datasets.
(C) It simplifies the computation process.
(D) It eliminates the need for validation data.

8. MAE is less sensitive to outliers than MSE or RMSE because it does not square the errors. What does an MAE of 0 indicate?
(A) The model has high variance.
(B) The model has a high bias.
(C) The model’s predictions are perfect.
(D) The dataset contains no outliers.

9. Which metric would be the most appropriate for a regression model with a skewed distribution of errors?
(A) Mean Squared Error (MSE)
(B) Root Mean Squared Error (RMSE)
(C) Mean Absolute Error (MAE)
(D) R-squared

10. Which metric can show a value between -∞ and 1, where 1 indicates a perfect fit?
(A) Mean Absolute Error (MAE)
(B) Mean Squared Error (MSE)
© R-squared
(D) Root Mean Squared Error (RMSE)

The solutions will be published in the next blog ML Interview Prep 21. Happy learning. If you like the questions and enjoy taking the test, please subscribe to my email list for the latest ML questions, follow my Medium profile, and leave a clap for me. Feel free to discuss your thoughts on these questions in the comment section. Don’t forget to share the quiz link with your friends or LinkedIn connections. If you want to connect with me on LinkedIn: my LinkedIn profile.

The solution of the previous quiz — Regression Accuracy Metrics (Part 1): Machine Learning Interview Prep 19: 1(D), 2(D), 3(D), 4(B), 5(B), 6(A), 7(D), 8(C), 9(D), 10(A)

References:
[1] R-squared, Clearly Explained!!!
[2] How to Calculate Accuracy for Regression?
[3] Beyond Accuracy: Understanding the Performance Metrics of Regression Models
[4] Regression Metrics

--

--