Every article in our adverse media database receives a relevancy score between 50 and 100. This score determines which results appear when you run an adverse media search and helps ensure the most relevant articles surface first.
Relevancy scoring combines four factors:
Name Match
Our system compares extracted perpetrator and victim names against your search term using a modified Jgram algorithm. Perpetrator name matches receive higher scores than victims, to ensure that you are seeing relevant results for searches.
Relevancy Weightings
Articles receive score boosts based on:
- Confidence: How certain we are about the article's contents
- Sentiment: Whether the tone is negative
- Severity: The seriousness of the allegations or incidents
Global Rank
Major news sources and government sites receive score boosts over more obscure sources like blogs and content aggregators.
Language
English articles receive a slight boost while non-English articles get a small penalty. The minimum score threshold is adjusted so relevant non-English results still appear.
How Are Results Sorted?
Results use a three-tier sorting system:
- Score (primary): Highest scores appear first
- Global Rank (tiebreaker): Higher-ranked sources appear first if scores match
- Published Date (final tiebreaker): More recent articles appear first if score and rank match
Each tier only applies when the previous levels are tied.