Bookmobile arrival errors are consistently larger for one neighborhood even though the overall RMSE looks fine. What should the analysis notice?
Select an answer to reveal the explanation.
Short Explanation
Bookmobile errors are fat in one neighborhood while overall RMSE looks fine. That is prediction bias: an aggregate metric can hide a harmed subgroup. Polly is not a subgroup RMSE, and a neighborhood gap is not an S3 throughput issue.
Full Explanation
Detect model (prediction) bias when an aggregate metric hides a harmed subgroup. A fine overall RMSE can still leave one neighborhood worse. Polly is not that analysis, and a neighborhood gap is not an ingest-capacity issue.