统计研究 ›› 2010, Vol. 27 ›› Issue (6): 100-105.

• 论文 • 上一篇    

Bootstrap方法在死亡模型中的应用

孙佳美 段白鸽   

  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2010-06-15 发布日期:2010-06-15

The Application of Bootstrap Method in Mortality Model

Sun Jiamei & Duan Baige   

  • Received:1900-01-01 Revised:1900-01-01 Online:2010-06-15 Published:2010-06-15

摘要: 由于不同国家死亡率改善现象不同,世界各国所使用的死亡率模型皆不尽相同,而且不同年龄段的死亡率模型也不同。实际中,我们常常采用Gompertz模型、Makeham模型、Weibull模型等拟合高年龄段人口的死亡率,但是因高年龄段人口的死亡数据资料不够充分,较少有人以统计的观点给出模型适合性的检验过程。因此本研究提出利用Bootstrap方法检验死亡模型假设的方法,包括模型适合性的检验、参数估计、参数假设检验等。最后,本文应用中国1997-2007年65-89岁人口的粗死亡率数据,提出适合的死亡模型,然后给出利用Bootstrap方法进行死亡模型检验的全过程。

关键词: Bootstrap方法, 死亡模型, 参数估计, 假设检验

Abstract: Mortality rate models are not the same due to different countries and different ages. We often use Gompertz model, Makeham model, Weibull model to fit the mortality data of old age groups. Because the mortality data of old age groups is often insufficient, no one tests the fitness of the selected mortality model. Therefore, in this paper, Bootstrap methods are used to test the assumptions of mortality model, including the fitness of the mortality model, parameter estimation, parametric hypothesis testing. Finally, using the crude death rate of China population aged 65 to 89 from 1997 to 2007; we establish the mortality model and give the whole process of model checking by applying apply Bootstrap method.

Key words: Bootstrap, Method, Mortality, Model, Parameter, Estimation, Hypothesis, Testing