摘要
未决赔款准备金评估的随机性方法逐渐得到重视,而考虑赔款数据的相关性可提高准备金评估的精确性。在确定性期望赔付法的基础上,提出一种基于阿基米德Copula函数的随机性期望赔付法;在准备金评估中利用核密度估计实现进展因子的随机化,并在此基础上应用阿基米德Copula函数分析两类赔款数据相关性的问题;利用R软件模拟总损失准备金的分布,研究表明该方法相比传统的期望赔付法具有更强的灵活性,其结果也更符合实际。
The stochastic approach to the assessment of outstanding claims reserves is increasingly taken into account, and the relevance of the data can improve the accuracy of the reserve assessment. On the basis of deterministic expectation payment method,a stochastic expectation payment method based on Archimedean copulas is proposed. Using the kernel density estimation to realize the randomization of the progress factors in the reserve evaluation. And apply Archimedean copulas to analyze the correlation of the two types of claims data. And then use the R software for empirical analysis to simulate the distribution of the total reserves. The results show that the proposed method is more flexible than traditional expectation method and the result is more realistic.
出处
《统计与信息论坛》
CSSCI
北大核心
2017年第8期8-15,共8页
Journal of Statistics and Information
基金
国家自然科学基金项目<基于结构化大数据深度挖掘的非寿险保险公司经营风险模型研究>(61502280)