Link analysis for fraud detection

How our technologies help to detect and investigate fraud

The anti-fraud data challenge

Fraud is an expensive and complicated problem. To understand and detect it, you need to see connections – between people, accounts, transactions, and dates.

It’s further complicated by the fast-changing tactics used by fraudsters. As their crimes become more sophisticated, so too must the data analysts’ tools. This means relying on powerful link analysis.

Webinar: Fighting Fraud with Graphs

The link analysis solution

When trying to detect fraud, analysts are looking for two different kinds: known and unknown fraud.

Fraud investigation and detection techniques are used to uncover known and unknown fraud
Fraud investigation and detection techniques are used to uncover known and unknown fraud

As the names suggest, ‘known fraud’ is fraudulent activity that has been encountered before. The behavior patterns can be defined and detected using rule-scoring and pattern-matching algorithms. In this case, link analysis plays an investigative role, providing fast and accurate situational analysis and informing a go / no-go decision.

Conversely, ‘unknown fraud’ has not been previously encountered. As automated processes will not help here, link analysis acts as a detection tool.

Investigations of known fraud tend to take a case-centric approach – starting from a specific transaction, account or person, and working outwards. We can see that approach in action in this blog post about insurance fraud.

Unusual connections in an insurance fraud investigation
Looking for known fraud patterns in an insurance fraud link analysis example

Investigations of unknown fraud use a global approach – loading large volumes of transactions into a chart to uncover outliers or unusual patterns.

Using a global approach to uncover unusual patterns that could indicate previously 'unknown' fraud
Using a global approach to uncover unusual patterns that could indicate previously ‘unknown’ fraud

White paper: fraud detection & investigation

Discover five ways our customers use link analysis to detect and investigate fraud.

Download the White Paper

Why choose us?

Fraud data is large, complex, noisy and often incomplete. Link analysis helps analysts investigate fraud in an interactive and intuitive way.

Our technologies can help analysts:

  • Join the dots in fraud data to discover patterns and anomalies
  • Pull transaction data from multiple sources
  • Share interactive charts for reporting and investigation
  • Make informed decisions more quickly

Our toolkits also include advanced functionality to help analysts unlock fraud data and see context on demand, including filtering, layouts, social network analysis, time analysis and geospatial visualization.

Find out more

We work with organizations, including Visa, Aviva, Cifas and JPMC, to help them make sense of complex connected data for fraud management. If you’d like to learn more, request a trial of our technology.

Try our technology