การพยากรณ์เบี้ยประกันภัยแท้จริงของการประกันภัยรถยนต์ โดยใช้ทฤษฎีความน่าเชื่อถือผ่านแนวคิดของที-คอปปูลา
The purpose of this study is to forecast the pure premium for automobile insurance (only for the own damage) using t-copula credibility theory and Buhlman-Straub model. The mean squared error is used to compare the results from these two methods. Data used in the study are the own damage claims duri...
Saved in:
| 主要作者: | |
|---|---|
| 其他作者: | |
| 格式: | Theses and Dissertations |
| 語言: | Thai |
| 出版: |
จุฬาลงกรณ์มหาวิทยาลัย
2011
|
| 主題: | |
| 在線閱讀: | https://digiverse.chula.ac.th/Info/item/dc:28232 |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 機構: | Chulalongkorn University |
| 語言: | Thai |
| 總結: | The purpose of this study is to forecast the pure premium for automobile insurance (only for the own damage) using t-copula credibility theory and Buhlman-Straub model. The mean squared error is used to compare the results from these two methods. Data used in the study are the own damage claims during 2006 to 2009 for comprehensive type, private cars (code 110), engine size not more than 2,000 c.c., vehicle group 3, vehicle age 2-5 years and driver age 24-35 years old. These data sets were classified by 11 regions and 4 vehicle age groups. This study showed that the forecasting pure premiums using the Buhlman-Straub model are higher by the vehicle age. In this study, the t-copula credibility theory can be applied for the claim data of the vehicle age 2 years only due to the assumption of the dependencies over time. Two explanatory variables, registered car densities and population, were identified as being related to the own damage claims. The mean squared error of t-copula credibility method is less than that of Buhlman-Straub predictors. |
|---|