Databricks Machine Learning Associate Question 611
Single answerYou are building a machine learning model to predict house prices. During preprocessing, you apply a log transformation to the target variable (house price) to stabilize variance and improve model performance. After training the model, you calculate the RMSE (Root Mean Square Error) on the log-transformed scale. What must you do to correctly interpret the RMSE in terms of the original house prices?
- A
Exponentiate the RMSE to reverse the log transformation and interpret it in the original scale.
- B
Directly interpret the RMSE as it is, since the log transformation does not affect RMSE interpretation.
- C
Exponentiate the predictions and ground truth values, then recalculate the RMSE to interpret it in the original scale.
- D
Exponentiate the RMSE and subtract 1 to reverse the log transformation and interpret it in the original scale.
Show answer and explanation
Correct answer: C
Explanation
When working with log-transformed variables, evaluation metrics like RMSE that are calculated on the transformed scale cannot be directly interpreted in the original scale. To interpret them correctly, you need to exponentiate the predictions and ground truth values back to the original scale and then recalculate the RMSE. This ensures that the metric reflects the actual errors in the original units of the target variable (e.g., house prices).
- A. Incorrect.
This is incorrect because RMSE is not directly exponentiated. The RMSE must be recalculated after exponentiating the predictions and ground truth values to properly interpret it in the original scale.
- B. Incorrect.
This is incorrect because the RMSE calculated on the log-transformed scale cannot be directly interpreted in terms of the original scale. Log transformations change the numerical meaning of the values.
- C. Correct.
This is correct because RMSE should be recalculated after exponentiating the predictions and ground truth values back to the original scale. This ensures the metric is meaningful in terms of the original house prices.
- D. Incorrect.
This is incorrect because subtracting 1 after exponentiating the RMSE is not a valid mathematical operation for reversing the log transformation.