問題文
A team needs a dataset that behaves like production so that it can exercise a new fraud detection rule, but the institution does not allow real customer values in the test environment under any approval route. Which approach fits this constraint?
選択肢
- Duplicate the operational counterparties so that the replicated dataset keeps the original identifiers.
- Truncate the retained chronology so that the diminished dataset retains the preceding transactions.
- Encipher the delivered statements so that the protected dataset keeps the sealed attachments.
- Generate the synthetic transactions so that the produced dataset keeps the real patterns.