A Research On Recognition Model Of Automobile Insurance Fraud

With the development and improvement of our national strength and nationaleconomy level，it shows that，the possession of motor vehicles per capita is in a rapidgrowth trend. As it’s shown by the National Bureau of Statistics in2012, thepopulation late2012has reached to120,890,000, an increase of14.3%from the endof the previous year. And this development brings us the rapid growth of the motorvehicle accidents in China. To avoid the risk, people choose to buy vehicle insurance.But the increase of the insurance didn’t bring the insurance companies huge profits.Actually, there are many insurance companies suffering a big loss, the mostimportant reason is the proliferation of motor vehicle insurance fraud.Response to this phenomenon, we need a relatively reasonable andcomprehensive motor vehicle insurance fraud identification system. On the eye ofInsurance itself, it should establish a strict underwriting system, and focus ontraining the professionals to a high quality. On the eye of law, the legal system of ourcountry is not very sound, so specific laws and constitution should be developed andspecialized anti-fraud agencies should be established. From the social aspect, peopleusually hold an understanding and tolerant attitude to motor vehicle insurance fraud.This has helped insurance fraud of motor vehicle, so social credit system should beestablished to limit the development of the situation. At the same time, inMathematical, we should focus on the study of developing effective recognitionmodel to deal with the insurance fraud.In this paper, I summarized the application of mathematical models in the past,which are from scholars of motor vehicle insurance fraud recognition. Based onthem and statistical theory, introduced the Generalized Linear Model. Study its properties and applied to the identification of the motor vehicle insurance fraud anddo the parameter estimation and data fitting. Aims to get a recognition model betterever.