[1]鲁亚会,刘爱义.基于半连续两部模型的保险损失预测[J].浙江科技学院学报,2023,(06):467-474.[doi:10.3969/j.issn.1671-8798.2023.06.002 ]
 LU Yahui,LIU Aiyi.Prediction of insurance loss based on semicontinuous two-part model[J].,2023,(06):467-474.[doi:10.3969/j.issn.1671-8798.2023.06.002 ]
点击复制

基于半连续两部模型的保险损失预测(/HTML)
分享到:

《浙江科技学院学报》[ISSN:1001-3733/CN:61-1062/R]

卷:
期数:
2023年06期
页码:
467-474
栏目:
出版日期:
2024-01-05

文章信息/Info

Title:
Prediction of insurance loss based on semicontinuous two-part model
文章编号:
1671-8798(2023)06-0467-08
作者:
鲁亚会1刘爱义2
(1.浙江科技学院 经济与管理学院,杭州 310023; 2.美国国立卫生研究院,美国 贝塞斯达 20817)
Author(s):
LU Yahui1 LIU Aiyi2
(1.School of Economics and Management, Zhejiang University of Science and Technology, Hangzhou 310023, Zhejiang, China; 2.National Institutes of Health, Bethesda 20817, Maryland, USA)
关键词:
累积损失预测 半连续数据 Tweedie回归模型 两部回归模型
分类号:
F842; O212.1
DOI:
10.3969/j.issn.1671-8798.2023.06.002
文献标志码:
A
摘要:
【目的】提高保险领域中保单累积损失预测的准确率。传统的Tweedie回归模型只能对非零均值建立回归模型,却不能对零概率建立回归模型,从而导致该模型的拟合效果并不理想。【方法】考虑到保单损失数据中往往包含着大量的零索赔,此时可视其为一种半连续型数据。因此,基于半连续两部模型,并考虑到累积损失中非零连续部分的分布类型,提出3种不同的累积损失预测模型,并结合一组实际损失数据进行模型对比分析。【结果】与Tweedie回归模型相比,本研究所提出的半连续两部回归模型的赤池信息准则值(Akaike information criterion,AIC)和贝叶斯信息量准则值(Bayesian information criterion,BIC)更小,具有较好的拟合效果。【结论】本研究结果可为保险领域中的保单累积损失预测提供参考。

参考文献/References:

[1] BIGNOZZI V, PUCCETTI G, RUSCHENDORF L. Reducing model risk via positive and negative dependence assumptions[J].Insurance Mathematics & Economics,2015,61(1):17.
[2] JORGENSEN B, PAES D S. Fitting Tweedie's compound Poisson model to insurance claims data[J].Scandinavian Actuarial Journal,1994,1994(1):70.
[3] SMYTH G K. Fitting Tweedie's compound Poisson model to insurance claims data:dispersion modelling[J].Astin Bulletin,2002,32(1):145.
[4] 孟生旺,李政宵.基于随机效应零调整回归模型的保险损失预测[J].统计与信息论坛,2015,30(12):7.
[5] MADDEN D. Sample selection versus two-part models revisited:the case of female smoking and drinking[J].Journal of Health Economics,2008,27(2):303.
[6] OLSEN M K, SCHAFER J L. A two-part random-effects model for semicontinuous longitudinal data[J].Journal of the American Statistical Association,2001,96(454):732.
[7] SU L, TOM B D M, FAREWELL V T. Bias in 2-part mixed models for longitudinal semicontinuous data[J].Biostatistics,2009,10(2):378.
[8] BOCK H M, BRENNER H. Inequalities in out of pocket payments for health care services among elderly Germans:results of a population-based cross-sectional study[J].International Journal for Equity Health,2014,13(1):3.
[9] RUSTAND D, BRIOLLAISI L, TOURNIGAND C, et al. Two-part joint model for a longitudinal semicontinuous marker and a terminal event with application to metastatic colorectal cancer data[J].Biostatistics,2022,23(1):50.
[10] 鲁亚会,刘爱义,江涛.多变量半连续数据的似然比检验[J].系统科学与数学,2021,41(11):3258.
[11] OLSEN M K, SCHAFER J L. A two-part random-effects model for semicontinuous longitudinal data[J].Journal of the American Statistical Association,2001,96(454),730.
[12] JAFFA M A, GEBREGZIABHER M, JAFFA A A. Shared parameter and copula models for analysis of semicontinuous longitudinal data with nonrandom dropout and informative censoring[J].Statistical Methods in Medical Research,2022,31(3):468.
[13] 段星德,张实,罗露璐,等.Tweedie复合泊松回归模型的Bayes估计和影响分析[J].高校应用数学学报A辑,2020,35(4):393.
[14] 李康,史宪铭,李广宁,等.基于正态-逆伽马分布的反巡航导弹命中概率估计方法[J].系统工程与电子技术,2022,44(8):2621.
[15] 赵远英,徐登可,冉庆.泊松逆高斯回归模型的贝叶斯统计推断[J].应用数学,2021,34(2):253.
[16] WANG T, CHENG J, GENG J. Wave equation reflection traveltime inversion using Gauss-Newton optimization[J].IEEE Geoscience and Remote Sensing Letters,2022,19:19.
[17] 韦博成.参数统计教程[M].北京:高等教育出版社,2006.
[18] CHARPENTIER A. Computational actuarial science with R[M].London:CRC Press,2014.

备注/Memo

备注/Memo:
收稿日期:2022-11-19
基金项目:杭州市哲学社会科学规划课题(Z23JC042); 国家自然科学基金项目(11971433)
通信作者:鲁亚会(1990— ),女,河南省商丘人,讲师,博士,主要从事应用统计研究。E-mail:luyahui92@163.com。
更新日期/Last Update: 2023-12-31