基于LogGED-GPD模型的巨灾损失分布拟合  

Distribution of Catastrophe Loss Distribution Based on LogGED-GPD Model

作　　者：王永茂[1] 杨晓婷 WANG Yongmao;YANG Xiaoting(College of Science,Yanshan University,Qinhuangdao 066004,China)

机构地区：[1]燕山大学理学院,河北秦皇岛066004

出　　处：《郑州大学学报（理学版）》2021年第3期100-104,共5页Journal of Zhengzhou University:Natural Science Edition

基　　金：廊坊市科技局科学技术研究项目(2016011031)。

摘　　要：近年关于巨灾损失分布模型的研究方法多采用单一分布模型,或具有固定权重的组合分布模型。在对数广义误差分布(LogGED)的基础上,运用广义帕累托分布(GPD)拟合数据的厚尾部分,并加入可变权重组合分布模型的拟合思路,构建了可变权重的对数广义误差-广义帕累托组合分布模型(LogGED-GPD)。并对全球洪水巨灾损失进行了实证分布拟合,验证了可变权重的对数广义误差-广义帕累托组合分布模型的拟合效果更好,一定程度上为今后洪水巨灾损失的分析方法提供了参考依据。Based on the logarithmic generalized error distribution(LogGED),the generalized Pareto distribution(GPD)was used to fit the thick tail part of the data,and the fitting idea of the variable weight combination distribution model was added to construct the logarithmic generalized error and generalized Pareto combination distribution model(LogGED-GPD).The empirical distribution of global flood catastrophe losses was carried out,and it was verified that the logarithmic generalized error-variable Pareto combination distribution model with variable weights had a better fitting effect,and to some extent,the analysis of flood catastrophe losses in the future.

关 键 词：对数广义误差分布 广义帕累托分布 组合分布模型 全球洪水巨灾损失 

分 类 号：F840[经济管理—保险]