Question 1

How does name matching work in screening?

Accepted Answer

Modern screening engines combine multiple matching techniques: exact match, normalised exact (case, accents, stop-words removed), phonetic algorithms (Soundex, Metaphone), edit-distance (Levenshtein, Damerau-Levenshtein), token-based similarity (Jaccard, Jaro-Winkler), transliteration across scri...

Question 2

What is a false positive?

Accepted Answer

An alert raised by a screening engine where the matched record is not in fact the same person or entity as the customer. False positives are unavoidable: the cost of a missed true hit is so high that engines err toward over-matching. Industry norms vary widely but a typical sanctions-screening fa...

Question 3

What is a false negative?

Accepted Answer

A true hit that the engine fails to raise — the screening run completes "clean" but the customer is in fact a sanctioned or PEP individual. False negatives are far more dangerous than false positives because they expose the institution to legal liability. They are typically caused by aggressive d...

Question 4

What is alert disposition?

Accepted Answer

The documented outcome of a human (or automated) review of a screening alert: typically one of True Match, False Positive, or Pending / Escalated. Each disposition must be supported by the reviewer's reasoning and the evidence consulted. Disposition data is itself a valuable input for tuning thre...

Question 5

Why is governance of screening so important to regulators?

Accepted Answer

Because screening is the primary technical control against the most serious financial-crime offences, regulators expect demonstrable governance: model risk management, periodic threshold tuning with statistical validation, independent testing, change control on rule sets and reference data, audit...

Screening mechanics: matching, false positives, governance