Summary: | 碩士 === 國立高雄第一科技大學 === 風險管理與保險研究所 === 104 === With social progress, people are suffering from accidental injury and hence the number of such cases is increasing. As such events happen and benefit are claimed, the insured will easily dispute with their insurance companies, so ensuing litigation problems follow and the time to deal with claims for the insurance company is prolonged.
In this study, I use the R Language to discuss whether the level of injury, the time to deal with claims for insurance companies, and the occurrence of litigation problems affect the amount of claims; the linear regression model and Gamma regression model are used to test those effects. Since linear regression model is a special case of generalized linear model, the normal regression model, therefore, is chosen when it is compared with gamma regression model by Akaike’s information criterion (AIC). Ggplot2, which is considered to be one of the best plotting packages, is used to visualize the data, and the package rgl is used to establish a 3D model.
The empirical results show that the AIC of the normal regression is relatively low when it is compared to those of other models. Therefore, the normal regression model is chosen the final model, along with the findings which show that all factors of the level of injury, the occurrence of litigation problems, and the time to deal with claims for the insurance companies have significant effects on the amount of claims, which mean that the injury level, hiring a consultant or not, and the prolonging of the time to deal with the accident will affect the amount of claims respectively
|