Skip to Content
TechnologyStrise Knowledge Graph

Strise Knowledge Graph

Connecting the Business World

The Strise Knowledge Graph is a dynamic, constantly evolving network designed to help understand the complex web of relationships in the business landscape.

What is it

At its heart, the Knowledge Graph is a vast database structured like a spiderweb with two fundamental components:

  • Entities: Companies, Persons, Locations, Industries, and Topics — the key players in the business world.
  • Relationships: The connections showing how entities are linked — from board positions to ownership to geographic operations.

What data is in it

Entities

  • Companies: Ownership structures, board members, financial information, industry classifications
  • Persons: Beneficial owners (UBOs), board members, stakeholders, with metadata like nationality and date of birth
  • Locations: Operational addresses, headquarters, geographic presence
  • Industries: Categorizations for targeted risk analysis
  • Topics: Themes like adverse media, high risk, money laundering

Relationships

  • Ownership: Direct and indirect ownership structures, including UBOs
  • Board Memberships: Individuals connected to their roles within companies
  • Regulatory Links: Ties to sanctions, PEPs, and adverse media
  • Geographic Connections: Links to high-risk or sanctioned regions

Events and Updates

  • Media Mentions: Articles enriched with relevance scoring
  • Regulatory Changes: Sanctions, PEP statuses, compliance events
  • Corporate Changes: Registry updates on ownership, structure, or financials

How it works

Data ingestion

The system collects vast amounts of data from diverse sources, normalizing and standardizing it for the graph’s structure.

Contextual understanding

NLP disambiguates terms — distinguishing “Apple” the company from “apple” the fruit by analyzing context.

Entity linking

An advanced Entity Linking (EL) module connects text in documents to entities in the Knowledge Graph:

  • Candidate Selection: Identifies potential entity matches
  • Graph-Based Scoring: Uses PageRank-inspired techniques to score entities
  • Machine Learning: Refines matches using popularity and frequency features

Graph integration

Validated entities are integrated by creating or updating nodes (entities) and edges (relationships).

Revealing insights

The graph traverses relationships to detect indirect ties — for example, a flagged owner’s connections to a high-risk country.

Value proposition

  • Holistic view: See how facts are interconnected, not just isolated data points
  • Unbiased perspective: Diverse sources provide balanced understanding
  • Hidden relationships: Identify risks that would be missed in individual reports
  • Actionable intelligence: Know if an owner has been sanctioned, even without direct mention in any single source

Key Takeaways

The Knowledge Graph is:

  • A constantly evolving network mirroring the dynamics of the business world
  • A system that continuously learns from new data
  • A tool that connects the dots to reveal hidden relationships
  • A way to turn complex, raw data into actionable insights
Last updated on