| 摘要 |
<p>A system (Fig. 1) and method of detecting of insurance premium fraud due to the misrepresentation of policy related information by the policyholder is provided by a predictive model (Fig. 2), which uses peer group policy data (205), historical variables (203), and internal variables (207) to assess the likelihood of premium fraud for each policy. The predictive model (Fig. 2) produces a score (410) for the policy, which is a measure of the likelihood of a misrepresentation of policy information. The score is then ranked, and the policy is audited to confirm premium fraud or abuse.</p> |
