Home ] Solutions ] Technology ] Demos ] Downloads ] About Us ] Contact Us ]
  Fraud Detection
Matching solutions for fraud detection.

Healthcare
Insurance
Needs Analysis
Fraud Detection
Contract Generation
e-Procurement
e-Government
e-Markets

 

 

Behaviour pattern matching

The Matching Systems Engine is an ideal runtime vehicle for detecting fraud.  Known behaviour patterns, both fraudulent and benign, make up the set of advertisements that are held in the Matching Systems Engine.  These can be simple static patterns against which a target customer behaviour will be matched.  The resulting matching environment can then be embedded in an existing java application, or made available as a Web Service to multiple systems.  It will automatically find known behaviour patterns that match a given target customer behaviour.  Scoring algorithms can be applied to those that have matched to rank them in order of proximity.

The power of dynamic patterns

With the Matching Systems Engine they can also be dynamic patterns, whose content changes depending on the individual target customer behaviour patterns that are presented.  For example, a fraudulent spending pattern may relate not to absolute spending thresholds, but to ones that have been derived from that particular customer's typical spending behaviour.  This kind of dynamic pattern is very easy to specify with the Matching Systems Engine.  It can involve dynamic calculations using the target customer behaviour information, and if needed, dynamic access to external information systems.

Navigating rich information

Unlike most rule engine technologies, the Matching Systems Engine was designed with complex information structures in mind.  XPath expressions can be used to navigate target customer behaviour information, facilitating the easy construction of a rich pattern matching environment with complex information structures.

High performance matching

Fraud detection is a critical activity, often requiring deployment in a high-performance real-time environment.  The Matching Systems Engine does not rely on a rule interpreter, with all its inherent performance and management difficulties.  Advertised behaviour patterns are automatically converted into Java classes, enabling them to be assessed with maximum efficiency.

Keeping one step ahead

It is vital that you have instant control over your fraud detection system.  With Matching Systems Engine, behaviour patterns can be withdrawn or amended, or new ones added, in an instant.

 

 

 © Matching Systems 2007 Ltd., July 2007