Question 410
An AutoML classification experiment includes rows from the same hospital patients across several visits. A random row split lets future visits from a patient appear during training while an earlier visit is in validation. What holdout strategy should be used?
Explanation
Correct Answer
Entity-aware and temporal partitioning can prevent record-level leakage that a random row split misses.
Incorrect Answers
Seed changes may alter the leak distribution but do not enforce independent patients or chronology.
Collapsing identifiers destroys the grouping information needed to prevent leakage.
Inference batching occurs after training and has no effect on holdout construction.