Databricks Machine Learning Associate Question 601
Select 2You are tasked with evaluating the performance of a regression model built to predict house prices. The model predictions and true values are provided, and you compute the following metrics: RMSE = 25, MAE = 20, and R-squared = 0.85. Based on the metrics, which of the following statements accurately describe the model's performance?
- A
The model's predictions are, on average, 25 units away from the true values.
- B
The model captures 85% of the variance in the target variable.
- C
The model's predictions have a mean absolute error of 25 units.
- D
The RMSE being higher than the MAE indicates the presence of large prediction errors.
- E
An R-squared value of 0.85 means the model is underfitting the data.
Show answer and explanation
Correct answers: B, D
Explanation
To evaluate regression models, metrics like RMSE, MAE, and R-squared provide insights into the model's error and explanatory power. A high R-squared value (0.85 here) indicates that the model explains a significant portion of the variance in the target variable. Additionally, the relationship between RMSE and MAE can indicate the presence of large errors, as RMSE penalizes larger deviations more than MAE.
- A. Incorrect.
This statement is incorrect because the RMSE (Root Mean Squared Error) represents the square root of the average squared differences between predicted and actual values, not the average absolute difference.
- B. Correct.
This statement is correct. R-squared (coefficient of determination) quantifies the proportion of the variance in the target variable explained by the model. A value of 0.85 indicates the model captures 85% of the variance.
- C. Incorrect.
This statement is incorrect because the Mean Absolute Error (MAE) is explicitly provided as 20, not 25.
- D. Correct.
This statement is correct. RMSE being higher than MAE suggests the presence of larger errors in the predictions, as RMSE penalizes larger errors more heavily due to squaring.
- E. Incorrect.
This statement is incorrect. An R-squared value of 0.85 indicates good performance and does not imply underfitting. Underfitting typically results in a low R-squared value.