引導式測驗 · 硬 等級
利用 NLP 進行負面新聞篩選測驗
使用 NLP 來定位不良媒體並確定其優先順序,同時在決策前驗證身分、來源品質、相關性和人工審核。
問題 1 的 8
What should an NLP adverse-media alert represent?
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。
- What should an NLP adverse-media alert represent?
- A lead for a reviewer to investigate
- Proof the customer committed a crime
- An automatic reason to close an account
- A confirmed sanctions designation
- Why combine name matching with additional identifiers?
- To make every article legally authoritative
- To reduce false matches between people or organizations with similar names
- To remove the need for source review
- To ensure every alert is current
- What should a reviewer consider besides whether an article contains a customer name?
- 僅記錄模型的置信分數
- The number of words in the article
- Source reliability, context, recency and materiality
- Whether the article uses a negative adjective
- How does adverse-media screening differ from sanctions screening?
- It is always a government-issued list
- It proves whether a person is politically exposed
- It is only a way to detect spelling errors
- It examines public reporting for risk context rather than matching an official sanctions list
- What does a classifier score establish by itself?
- A prioritization signal under the model’s setup
- That an allegation is true
- That the right person was matched
- That the source is trustworthy
- Why retain the article URL, date and retrieval time?
- To replace the original investigation
- To preserve provenance and let a reviewer inspect the evidence
- To guarantee the publisher will not edit it
- To turn every article into a list match
- How should an institution handle an allegation in an article?
- Treat it as a proven fact
- Ignore it in every case
- Assess its credibility and relevance through the established review process
- Convert it directly into a sanctions match
- Which evaluation check can reveal unequal screening coverage?
- Measure only the average score
- Count alerts without checking results
- Use the same English source for every market
- Test coverage and errors across relevant languages and scripts