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133期讲座-Anastasios Panagiotelis,HanLi
(4月26日)
报告题目一：Forecast Reconciliation for Large Collections of Time Series
报告时间：2018年4月26日 下午2:00—3:00
报告人：
Dr Anastasios Panagiotelis is an Associate Professor at theDepartment of Econometrics and Business Statistics at Monash University.Anastasios recieved his PhD from the University of Sydney, and prior to joiningMonash in 2011, he was an Alexander von Humboldt Postdoctoral Researcher at theChair of Mathematical Statistics at the Technical University of Munich. Hisresearch interests include forecast reconciliation, Bayesian methodology andcopula models. He has been publishing in top statistics and econometricsjournals such as the Journal of American Statistical Association, Journal ofEconometrics and Journal of Business and Economics Statistics.
摘要：Largecollections of time series often have aggregation constraints, due togeographical or temporal hierarchies. When forecasts are made for each seriesindependently, it is rare for the forecasts to respect these aggregationconstraints in practice. Forecast reconciliation methods adjust forecastsand can be shown to improve forecasting performance. A review of forecastreconciliation methods will be given, including the least squares and Minimumtrace method. Two applications will be discussed, the first is anapplication to forecasting tourism demand in Australia, and the second isforecast admissions in a UK hospital.
报告题目二：Techniques to Analyze and Forecast Mortality
报告时间：2018年4月26日 下午3:00—4:00
报告人：
Dr Han Li is an AssistantProfessor at Macquarie University in Sydney, Australia. Before that, she workedat the University of New South Wales as a Senior Research Associate. Shereceived a Bachelor of Commerce (Honours) degree in Actuarial Studies at theUniversity of Melbourne and completed her PhD degree at Monash University. Shehas a broad range of research interests around longevity and mortality risks,population ageing and retirement financial products. Specifically, much of herresearch expertise is in the field of mortality modelling and forecasting usingadvanced econometric and statistical techniques. Han’s research has beenpublished in top tier journals including Insurance: Mathematics and Economicsand the Journal of Forecasting.
摘要：Theincreasing amount of attention paid on longevity risk and funding for old agehas created the need for precise mortality models and accurate mortalityforecasts. In this talk we are going to talk about some modern econometrics andstatistical techniques to better capture mortality patterns and improve theaccuracy of future mortality projections. These techniques will be appliedtomortality data from a wide range of developed countries including the GreatBritain, the United States, Australia, Netherlands, Japan, France and Spainover the post-war period 1950–2009. Contributions have been made to theexisting literature with focus given to the forecasting perspective of modelsand to the analysis of cohort effects.
报告地点：中央财经大学学术会堂南楼506（精算院会议室）
欢迎各位老师和同学积极参加！
