Free Databricks Machine Learning Associate Actual Exam Questions - Question 14 Discussion
this model, they have performed inference and the predictions and actual label values are in Spark
DataFrame preds_df.
They are using the following code block to evaluate the model:
regression_evaluator.setMetricName("rmse").evaluate(preds_df)
Which of the following changes should the data scientist make to evaluate the RMSE in a way that is
comparable with price?
A imo, exponentiating RMSE doesn’t fix scale issues, but converting predictions back first does.
A vs D? Exponentiating the RMSE itself (A) doesn’t really fix the scale mismatch because RMSE is a single error metric, not predictions. For a meaningful RMSE on price, you want both predictions and labels in the same scale before computing error. So transforming predictions back with exp (D) makes more sense. Just make sure actual labels are also on the original price scale, not logged, otherwise RMSE will still be off.
D imo, since exponentiating predictions aligns them with actual price for RMSE calculation.
D exponentiating the predictions makes sense because the model outputs log(price). This way, RMSE is calculated on the original price scale, matching the actual labels. Just make sure actual prices aren’t logged.
Maybe D makes the most sense because the model predicts log(price), so taking the exponent of predictions brings them back to price scale. If you just compute RMSE on log values, it’s not directly comparable to price. Options A and C seem off since RMSE isn’t just transformed by exponentiating after calculation or by sticking with log values. B is clearly wrong because predictions are already in log scale. So, transforming predictions before RMSE is the way to get a meaningful error measure in terms of actual price.
A exponentiating the RMSE is wrong because RMSE is not on the log scale itself.
Option D seems right since the model predicts log(price), so exponentiating predictions makes RMSE comparable to price.