•  

    Knowledge | Fraud Assessment | Traffic Accident

    This application demonstrates the use of an application deployed via a server environment.

    The REST based web service accepts JSON formatted data, executes the application logic and returns the results.

    In this particular demonstration, an AJAX call is made from the browser to the web service with the supplied data. The service could equally be called from any server environment which has the facility to make HTTP requests, for example .NET, PHP.

    The application contains natural language processing and pattern rules for fraud assessment. The application uses these features to determine whether a road traffic accident claim is fraudulant or not. If a claim is deemed to have a high risk of fraud, the claim is flagged for more closer scrutiny.

    In the demo when the default input values are used (Example 1) it causes a number of rules to be triggered (Example 2). In this particular situation one of the outcomes is ’roundabout, hotspot, multiple occupants’, in other words the accident occurred as a rear-end shunt at a roundabout in a hotspot area involving multiple claimants. This outcome appears as a result of the pattern rule highlighted (Example 3). The first condition ‘number_of_claimants > 3’ is satisfied as the number of claimants was 5, the second condition ‘accident_hotspot = TRUE’ is satisfied as the text “junction 1 of the M60” was found in the accident description text, the third condition ‘hit_in_rear_roundabout = TRUE’ is satisfied as the text “rear” and “roundabout” were both found in the accident description text (Example 4).

     

    RUN SAMPLE   MORE LIVE SAMPLES

     

    A schematic diagram of general transaction server architecture

     

    Schematic diagram of general transaction server architecture

     

    FA1

     

    Example 1. Default input values

     

    FA2

     

    Example 2. Result from default input values

     

    Fraud Assessment C Main Pattern Rules

     

    Example 3.  Main pattern rules

     

    Fraud Assessment D

     

    Example 4. Accident description text natural language processing

     

  • Try Now & Live Demos