摘要
分析广义线性模型和广义可加模型的理论基础和特点,从Tweedie类分布的独特视角归纳保险索赔额数据的分布规律。基于此建立了GLM-Tweedie和GAM-Tweedie索赔额拟合模型,以一组汽车保险损失数据为样本进行车险费率厘定和索赔额拟合的实证分析,识别"车、人、地"不同因素对费率不同的影响程度,助推我国车险费率厘定市场化改革精算技术的提升。
The paper analyzed the theory and characteristics of generalized linear model ( GLM ) and generalized addi- tive model (GAM) , and summarized the distribution of insurance claim from a unique perspective of Tweedie distribu- tions. It established GLM-Tweedie and GAM-Tweedie claim fitting models. Finally, based on a real dataset of automo- bile insurance loss, it made an empirical analysis of ratemaking and claim fitting, and identified the impacts of "vehicle, people, area" on rate to enhance the actuarial technique in the market-oriented reform of automobile insurance ratemaking in China.
出处
《天津商业大学学报》
2014年第1期60-67,共8页
Journal of Tianjin University of Commerce
基金
国家自然科学基金面上项目(71271121)
中央高校基本科研业务费专项资金"金融工程与精算学中的定量风险管理统计模型与方法"(NKZXTD1101)