A question about insurance fraud rule

[dincoo]

Hi everyone,

We have a rule that helps to find auto insurarance fraud and I am trying to code it.

The rule is: If the same vehicle_id's appear more than one incident.

The sample data is attached.

At the example there are two claims. First of them consists of 3 vehicles, second of them consists of 2 vehicles.

Thank you very much,

Onur

[EH]

Hello Onur,

I've added one suspect vehicle_id and prepared some code. This may give you a start...

data incidents;
   input INCIDENT_ID 1-7 VEHICLE_ORDER_NO 9 VEHICLE_ID $ 11-27;
   datalines;
4704002 1 JHMGD18707S205511 
4704002 2 VF1LAOF0520635569 
4704002 3 VF37ANFZWWP005871 
4704053 1 VF32BKFWA9D002173 
4704053 2 WF0HXXWPDH9C49160
4704053 2 VF1LAOF0520635569 
run;

proc sql;
   create table suspects as
      select vehicle_id, count(distinct incident_id) as nr_incid
        from incidents
    group by 1;

   create table incidents2 as
      select incidents.*, suspects.nr_incid
        from incidents, suspects
      where incidents.vehicle_id = suspects.vehicle_id;
quit;

proc print data=incidents2;
run;

Hope this helps,

Cheers, Eric