中国精算研究院

精算论坛第164期讲座 —彭亮 杨静平 张连增 (7月23日)

发布时间:2019-07-10 16:20    浏览次数:[]

龙马奋进 · 校庆70周年学术系列讲座

精算论坛第164讲座 彭亮 杨静平 张连增 (723)

教育部人文社科重点研究基地中央财经大学中国精算研究院学术活动

精算论坛第164讲座

(2019723)

 

 

报告题目一:

StatisticalInference for Mortality Model and Risk Allocation  

报告人: Liang Peng

Department of Risk Management andInsurance, Georgia State University

Peng isthe Thomas Bowles Chair professor in Actuarial Science in the Department ofRisk Management and Insurance at Georgia State University. His researchinterests include actuarial science, risk management, financial econometricsand statistics. Currently he is an associate editor of JASA, Statistica Sinicaand Scandinavian Journal of Statistics. Peng is fellow of the AmericanStatistical Association and fellow of the Institute of Mathematical Statistics.  

摘要:After revisitingthe well-known Lee-Carter mortality model in forecasting mortality risk andhedging longevity risk, we propose a modified Lee-Carter model to overcome thepossible inconsistent two-step inference proposed by Lee and Carter (1992).Unit root test, mortality forecast and bias corrected inference are provided too.Another topic is on risk or capital allocation, where a new allocation rulebased on Value-at-Risk is proposed and estimated nonparametrically at thestandard rate of convergence. In order to efficiently quantify the uncertainty,a residual based bootstrap method is proposed to combine with AR-GARCH models.


报告题目二:

Composite Bernstein Copula

报告人:杨静平教授(北京大学数学科学学院)

杨静平,北京大学数学科学学院教授,博士生导师。现任数量经济与数理金融教育部重点实验室(北京大学)副主任,中国工业与应用数学学会第七届理事会理事。研究兴趣有金融和保险中的风险相依性、债券组合模型和信贷资产证券化等。在金融数学期刊Finance and StochasticsSIAM Journal onFinancial MathematicsJournal of Computational Finance、精算学国际四大学术期刊以及概率论期刊Bernoulli和数学期刊Fuzzy Sets and Systems等发表了多篇学术论文。主持完成了中国国债发行策略的随机模拟模型、国债收益率曲线的拟合、信贷资产证券化以及含权债估值模型等方面的应用课题。  

摘要:Copula function has been widelyused in insurance and finance for modeling inter-dependency between risks.Inspired by the Bernstein copula (BC) put forward by Sancetta and Satchell(2004), we introduce a new class of multivariate copulas, the compositeBernstein copula (CBC), generated from a composition of two copulas. This newclass of copula functions is able to capture tail dependence, and it has areproduction property for the three important dependency structures:comonotonicity, countermonotonicity and independence. We introduce anestimation procedure based on the empirical composite Bernstein copula (ECBC)which incorporates both prior information and data into the estimation.Simulation studies and an empirical study on financial data illustrate theadvantages of the ECBC estimation method, especially in capturing taildependence. This is a joint work with Zhijin Chen, Fang Wang and Ruodu Wang.

报告题目三:  

神经网络在回归与分类问题中的应用  


报告人:张连增教授(南开大学金融学院)

张连增,理学博士。南开大学金融学院精算学系主任、教授、博导,中国精算师协会正会员。墨尔本大学、滑铁卢大学、洛桑大学等高校访问学者。研究方向主要包括精算风险理论、非寿险精算统计建模、机器学习精算应用等,多次受邀参加学术会议并作报告。已有20多年的高校教学科研工作经历,一直专注于精算教学与科研工作。主持国家自然科学基金和中国保监会课题,及校内高校基本科研业务项目等。代表性著作包括《寿险精算》(中国精算师资格考试指定用书),《未决赔款准备金评估的随机性模型与方法》,《精算学中的随机过程》等。部分研究论文发表于《保险研究》,《统计研究》,《数量经济技术经济研究》,《Insurance: Mathematics and Economics》,《NorthAmerican Actuarial Journal》,《Scandinavian ActuarialJournal》等期刊。长期任教精算风险理论、定量风险管理、保险统计分析等专业课程,教学经验丰富。另外,自始至终参与中国精算师资格考试工作。

摘要:本报告通过两个具体的实例,分别介绍神经网络在回归与分类问题中的应用。神经网络的训练和可视化,都可由R及软件包来实现。对回归问题,在评价神经网络效果时,可与通常的(广义)线性回归模型进行比较。对分类问题的效果评价,可与由其它方法(如回归模型、树方法)得到的ROC曲线及混淆矩阵加以比较。近年来,关于神经网络的车险定价精算应用已经成为研究热点。

 

 

报告时间:2019723 9:00--12:00


报告地点:学术会堂南楼506中国精算研究院会议室

 

欢迎各位老师和同学积极参加!