Databricks Machine Learning Associate Question 255
Single answerYou are training a regression model to predict house prices. During preprocessing, the log transformation is applied to the target variable (house prices) to handle skewed data. After training the model, you evaluate it using RMSE. Why might it be necessary to exponentiate the RMSE in this scenario?
- A
To convert the RMSE back to the original scale of the target variable.
- B
To improve the accuracy of the RMSE calculation.
- C
To remove the effects of overfitting in the model evaluation.
- D
To ensure that the RMSE is a positive value.
Show answer and explanation
Correct answer: A
Explanation
When the target variable is log-transformed during preprocessing, the RMSE calculated during evaluation will also be in the log scale. To interpret the RMSE in the original scale of the target variable (e.g., house prices), it is necessary to exponentiate the RMSE. This step ensures that the metric is meaningful and reflects errors in the same scale as the original data.
- A. Correct.
Exponentiating the RMSE converts it back to the original scale of the target variable (e.g., house prices). This is necessary because the RMSE is calculated based on the log-transformed labels, and without exponentiation, the RMSE remains in the log scale.
- B. Incorrect.
Exponentiating the RMSE does not improve its accuracy. The accuracy of RMSE depends on the model's predictions, not on whether it's exponentiated.
- C. Incorrect.
Exponentiating the RMSE does not address overfitting. Overfitting is related to how well the model generalizes to unseen data, not to the scale of the evaluation metric.
- D. Incorrect.
RMSE is inherently a positive value because it represents the square root of the mean squared error. Exponentiation is not required to ensure positivity.