WIP - Score Rolling Behaviour

Please respond to the below questions:

1. Has this content piece had a technical review by a TL (from a content accuracy and editorial perspective) Technical / Product content only
   • The editorial review must follow guidance from the Docs team, please use the resources linked at the bottom.
2. If part of the Tech Lead team - please use the tag 'Tech-Lead-Content'
3. Please specify the most suitable section and sub-section of the Community Library for publishing e.g. Service Operations: Operational Support
4. Who is your piece aimed at? (Data Engineer, Data Scientist, Investigator, Delivery Manager, etc.)
5. Can the article be linked to another piece/area of the Community?:
   1. A Specialist User Group (Data Management, FinCrime, KYC, or Insurance)
   2. Another related article in the Community
6. What is the purpose of your piece? Please include a summary line at the top of the blog.
7. Have you reviewed the content classification guidelines? Can the content be open to the public (i.e. searchable via Google) does it need to be restricted, or internal only?
   1. Please choose from the following:
      1. ✔️Public✔️- Content which is publicly viewable on the internet.
         1. Please note, the assumption is all content should be made public where possible due to the increased impact benefits
      2. 🔑Restricted🔑- Viewable by registered Customers and Partners, in addition to employees.
         1. For restricted content - Please provide a rationale for why this must be restricted.
         2. Create a short summary which can be posted in a public area of the Community - this short summary will link to the article and promote it.
      3. 🛑Internal🛑 - Viewable by Quantexa employees only.
8. Is there an urgency to publishing or preferential timeframe?
9. Editorial checklist:
   1. Have you capitalized any Quantexa terms like 'Entity', 'Document' and 'Entity Resolution'? (check the docs site Glossary if you aren't sure)
   2. Have you included a short intro paragraph at the top of the article, outlining the purpose and target audience, and a concluding paragraph including any relevant links to associated content?
   3. Have you used American (US) English - for example, 'analyze' not 'analyse'?
   4. Have you used headings, bullet points and tables to break up the content and grey boxes for code snippets?
10. Please include aliases, otherwise known as lines or alternate keywords for any terms you have used that are known by multiple names. This helps people find your piece even if they have not used the exact terms you've included. E.g. 'Elasticsearch (or Elastic)'
11. If you need further editorial guidance (beyond the guides below), tag James Parry or Ffion Owen.

Editorial Resources

Become a better writer ↗

Content models ↗

Quantexa writing guide ↗

Editing tools: Grammarly and Hemingway

This article covers rolling up and rolling down Scores. These concepts use the Score output from one level as input for another Score or Scorecard. Knowing these processes helps us build Scorecards. This approach allows us to capture risk accurately at all levels.

For context on what a Score means at each level (Document, Entity, or Network), refer to the Score Output Levels page.

This article is for Business Analysts, Business End Users, and Subject Matter Experts (SMEs). By grasping the concepts of rolling up and rolling down scores, they can make better decisions. This knowledge helps them aggregate risks from different data levels. These levels include Document, Entity, and Network.

Definition

Rolling Up

Rolling up a Score means aggregating Scores from a lower level (e.g., Document) to a higher level (e.g., Entity or Network).

This enables a wider risk assessment. It merges granular insights into a single, higher-level Score. For example, rolling up multiple transaction Scores to the Customer level gives a more fuller risk assessment. This approach incorporates all relevant transaction risks for that Customer.

Rolling Down

Rolling down a Score means distributing a Score from a higher level (e.g., Network or Entity) to a lower level (e.g., Document).

This ensures that risk assessed at a broader level is reflected at more granular levels. Rolling down a Network-level Score to all linked customers allows them to inherit relevant risks from their connections.

Rolling up and rolling down Scores helps group all Scores at the right level for Alerting

Why We Need To Roll Up or Down

Scores operate on different Quantexa objects: Documents, Entities, and Networks. In many cases, we need to analyze risk at different levels:

• When creating a Customer-centric Scorecard, you must combine several transaction-level (Document) Scores. This will give you the Business (Entity) level score for a Business Customer. This makes sure that all important transaction risks add to the total Business Customer risk.
• When we look at risk on a higher level, like the Network, we need to roll down the Network-level Score to all linked Customers. This way, we can give Customers the relevant risks based on their Network connections.
• To structure risk assessments well, you need to understand the output level for each Score and Scorecard.

Examples

Example of a Roll-Down Score

An example of a roll-down Score is one that aggregates Entity-level Scores to the Document level.

For example, a Customer Document may be linked to multiple Entities that trigger the same type of Score.

Suppose two individuals, John Hammerling and Norris Burke, are linked to the Customer Document for Turner & Page Book Shop Ltd. They both trigger the IndividualWithMultipleNationalIDs Score because they have several national IDs. We then introduce a new score: the IndividualWithMultipleNationalIDsRolldown Score. This score adds up the overall risk level of the Customer Document (Turner & Page Book Shop Ltd).

Figure 1: Example of a Roll-Down Score

The "Potential Identity Manipulation" Score adds up the Individual Entity Scores. The two Scores are aggregated to show the Document-level risk tied to the connected Entities.

Example of a Roll-Up Score

An example of a roll-up score is one that aggregates information from multiple Documents related to an Entity.

For example, an Account Entity may be scored by aggregating over the Customer Documents associated with it. The goal of rolling up is to create a more comprehensive Account Entity Score. It does this by using all the available Document-level data.

Another example would be a Business Entity linked to several Watchlist Documents. To determine if this Business Entity has Sanctions, we check each associated Watchlist Document. If a Watchlist Document contains relevant sanctions information, a Score is triggered. If no relevant information is found, the Score returns nothing and moves to the next Watchlist Document. After reviewing all Watchlist Documents, we use an aggregation function. This function creates an overall Entity Score using results from each Document.

Figure 2: Example of a Roll-Up Score

Where and How to Configure

[TBC]