引導式測驗 · 中等 等級

反洗錢交易監控:規則 vs. 機器學習測驗

測試基於風險的監控、規則與機器學習、警報含義、人工調查、模型驗證和報告職責。

問題 1 的 8

What does a transaction-monitoring alert establish?

預覽每個問題8 問題

當您參加測驗時,就會出現答案和解釋。

  1. What does a transaction-monitoring alert establish?
    • A confirmed suspicious activity report
    • Activity that may warrant investigation under the institution’s program
    • Proof a customer committed a crime
    • A required account closure
  2. 規則與機器學習監控有何不同?
    • Rules encode explicit patterns; ML can rank patterns learned from data
    • Rules never require tuning
    • ML always explains every alert fully
    • Rules prove intent while ML cannot use transactions
  3. What does FFIEC say may inform suspicious activity identification?
    • Only cash transactions
    • Only information from law enforcement
    • Employee observations, inquiries, advisories, and monitoring output
    • Only a machine-learning score
  4. How should an investigator handle a high-risk score?
    • Treat the score as proof of laundering
    • Ignore known customer information
    • Report the score without review
    • Review transaction and customer context before escalation
  5. Which result should model governance track?
    • Only total transaction count
    • Alert quality, workload, missed activity, and effects of changes
    • Only training runtime
    • 僅型號尺寸
  6. Why can ML monitoring fail after deployment?
    • Every model update improves detection
    • A model cannot be affected by labels
    • Transactions stop having context
    • Data definitions or risk patterns may shift over time
  7. What can rules contribute alongside ML?
    • A replacement for case investigation
    • A proof of criminal intent
    • Explicit controls for known patterns and requirements
    • A guarantee that no suspicious activity is missed
  8. Does sophisticated technology replace BSA/AML obligations?
    • No; the institution remains responsible for an effective risk-based program
    • Yes, if alert volumes fall
    • Yes, when the model is explainable
    • No, but staff can ignore alerts