统计研究 ›› 2013, Vol. 30 ›› Issue (9): 87-93.

• 论文 • 上一篇    下一篇

函数型死亡率预测模型

王洁丹等   

  • 出版日期:2013-09-15 发布日期:2013-09-04

Functional Model for Mortality Forecasting

Wang Jiedan et al.   

  • Online:2013-09-15 Published:2013-09-04

摘要: 人口死亡率反映人口的死亡水平,是人口规模的重要影响因素,同时也是人寿保险精算的重要数据基础。从数据特征来看,死亡率作为年龄的函数,是一种典型的函数型数据。本文使用函数型数据方法分析中国人口数据,基于1994年至2010年中国人口分年龄死亡数据,建立函数型死亡率预测模型,对未来分年龄死亡率进行预测,并通过生命表方法计算了未来平均预期寿命。同时通过对历史数据的预测,说明模型预测结果比较可信。

关键词: 函数型数据, 函数型预测模型, 死亡率预测, 函数型主成分分析

Abstract: Population mortality,reflecting the death level of the population,is an important factor of the scale of the population.At the same time,it is the significant data basis for life insurance actuarial science.On the characteristic of the data,as a function of age,Mortality is a typical functional data。This paper build a functional prediction model for mortality,based on Chinese age-specific mortality data from 1994~2010,forest the future aged-specific mortality and calculate the average life expectancy by the method of life table.In addition to this,prediction results by historical data shows that the prediction results is credible.

Key words: functional data, functional prediction model, mortality prediction, functional principal component analysis