培养工作

当前位置:研究生教育 >> 信息 >> 培养工作 >> 浏览文章

短课程:Actuarial Loss Modeling from a Data Analysis Perspective

发布时间:2017年05月08日 浏览次数: 文章作者:佚名 发布者:admin

短课程:Actuarial Loss Modeling from a Data Analysis Perspective

 

授课人:林小东 教授(加拿大多伦多大学)

 

课程简介:This course is on actuarial loss modeling for non-life insurance. It covers common severity models such as Gamma, Lognormal, Pareto and Weibull distributions. The teaching of those models and their applications will be done in an integrated manner. They are introduced one model at a time and I will go through the entire modeling process. I will begin with discussions on model properties, tail behavior, and its applications to insurance risk management, and move to simulation from the models and parameter estimations. Mixture models will then be introduced with an emphasis on model selection, estimation and validation. In particular, we will cover generalized finite Gaussian mixtures and finite Erlang mixtures, and provide EM algorithms for the mixture models.

 

参考书目:

1. Klugman,S. A.,Panjer, H. H.,&Willmot,G.E.(2012). LossModels:fromDatatoDecisions, 4th Edition. John Wiley & Sons.

2. Klugman,S. A.,Panjer, H. H.,&Willmot,G.E.(2013). LossModels:FurtherTopics. John

Wiley & Sons.

 

作业安排:There will be 2 assignments and students are required to write solutions in latex. There will also be 2 projects on numerical implementation and data fitting. Students are required to use R for all the numerical works and to write proper project reports in which the methods and approaches used in the projects are discussed and the findings are explained and interpreted.

 

选课要求:Only serious and highly motivated students should take this course. The coursework will be heavy so students who take this course should expect to work long hours.

 

上课时间:510日、11日、15日、17日、18日、22日,9-11

上课地点:数学物理大楼661