基于降水因子的湖南省财货险理赔风险指数模型  

Risk Index Model of Property and Cargo Insurance Claims Based on Precipitation in Hunan

作　　者：彭玮莹 黎跃勇 谭诗琪 李好 官建文 廖春花 周威 谢睿恒[1,3] Peng Weiying;Li Yueyong;Tan Shiqi;Li Hao;Guan Jianwen;Liao Chunhua;Zhou Wei;Xie Ruiheng(Hunan Provincial Meteorological Service Center,Changsha 410118;CAMS&China Re CRM·Joint Open Lab on Meteorological Risk and Insurance,Beijing 100081;Key Laboratory of Hunan Province for Meteorological Disaster Prevention and Mitigation,Changsha 410118;The People's Insurance Company(Group)of China Limited,Hunan Branch,Changsha 410118)

机构地区：[1]湖南省气象服务中心,长沙410118 [2]中国气象科学研究院&中再巨灾风险管理股份有限公司·气象风险与保险联合开放实验室,北京100081 [3]气象防灾减灾湖南省重点实验室,长沙410118 [4]中国人民财产保险股份有限公司湖南省分公司,长沙410118

出　　处：《应用气象学报》2025年第1期14-23,共10页Journal of Applied Meteorological Science

基　　金：气象风险与保险联合开放实验室开放基金项目(2023F007);湖南省气象局2021年重点科技项目(QXKJ21A006);中国气象局2024年度气象软科学研究课题(2024MSXM08)。

摘　　要：基于2016—2020年中国人民财产保险股份有限公司湖南省分公司因降水致灾的财货险理赔案例数据和湖南省降水数据,分析财货险中工程险、家财险和企财险的理赔案例与降水量间的关系,应用优势分析法构建湖南省不同财货险基于降水的理赔风险指数模型,计算风险指数等级阈值,输出理赔风险预警产品,并利用2021—2022年202个理赔案例进行模型检验。结果表明:降水因子中,24 h累积降水量、48 h累积降水量、72 h累积降水量和出险前1 d最大小时降水强度对理赔风险影响较大;各降水致灾因子对不同险种理赔风险的权重影响不同,48 h累积降水量影响最大,72 h累积降水量次之;202个案例中有146个案例检验准确,平均准确率达72.3%,轻度理赔风险等级准确率最高;个例检验表明未达到气象上的暴雨标准时,指数模型可以很好地对理赔风险进行预警。目前该指数模型已应用于湖南保险气象灾害风控服务平台中,基于智能网格预报数据生成分险种降水风险影响预报产品,对降水导致的财货险理赔风险进行预警服务。The relationship between insurance claims and precipitation levels is explored,especially in engineering insurance,home property insurance and enterprise property insurance within the realm of property and cargo insurance.It utilizes data from Hunan Branch of People's Insurance Company(Group)of China and Hunan Meteorological Station spanning from 2016 to 2020.By an advanced analysis method,a comprehensive risk index model is developed and tailored to different types of property and cargo insurance claims prevalent in Hunan Province.The primary objective is to assess how varying levels of precipitation affect the frequency and severity of insurance claims across these categories.Through the analysis of historical data,the key precipitation-related factors that significantly impact insurance claims are identified.These factors include 24-hour cumulative precipitation,48-hour cumulative precipitation,72-hour cumulative precipitation and the maximum hourly rainfall intensity preceding meteorological incidents.It is found that these factors exert varying degrees of influence on different types of property and cargo insurance.Notably,48-hour accumulated precipitation emerges as the most critical factor affecting claims,followed closely by 72-hour accumulated precipitation.These insights are derived from a rigorous analysis that integrates meteorological data with insurance claims data,enabling the development of a robust risk index model.Validation of the model's efficacy involves approximately 202 insurance claim cases spanning from 2021 to 2022.Results show a commendable accuracy rate of 72.3%,particularly strong in predicting instances of mild claims risk levels.Importantly,the model proves effective even in scenarios where meteorological rainstorm standards are not met,showcasing its reliability in preemptively identifying potential insurance claim risks.Presently,the developed risk index model has been implemented in Hunan Insurance Meteorological Disaster Risk Control Service Platform Version 2.0.The platform use

关 键 词：降水 财产保险 理赔风险 指数模型 

分 类 号：D92[政治法律—法学]