Longevity studies based on kernel hazard estimation

被引:13
作者
Felipe, A
Guillen, M
Nielsen, JP
机构
[1] Univ Barcelona, Dept Econ Estad & Econ Espanyola, Barcelona 08034, Spain
[2] Codan, Copenhagen, Denmark
关键词
counting process theory; kernel estimation; marker dependent hazards; boundary correction; mortality trends;
D O I
10.1016/S0167-6687(00)00076-7
中图分类号
F [经济];
学科分类号
02 [经济学];
摘要
We propose a non-parametric smoothing method to visualize the evolution of mortality rates. This is used to compare the mortality experiences of Denmark and Spain. Our comparison takes into account the dependence of mortality on chronological time and age. A two-dimensional surface indicating transition intensity is estimated and a number of other interesting estimators can be derived from this surface: i.e., ratios of the structural development of mortality, estimation of multiplicative structures using marginal integration (e.g., Linton and Nielsen Biometrika 82 [1995, 93-101]), and smoothed trends for fixed ages. The multiplicative structure effectively separates the chronological time component and the age component. (C) 2001 Elsevier Science B.V. All rights reserved.
引用
收藏
页码:191 / 204
页数:14
相关论文
共 29 条
[1]
Andersen P. K., 2012, Statistical models based on counting processes
[2]
ANDERSON JL, 1960, T 16 INT C ACT, V2, P379
[3]
[Anonymous], 1998, SCAND ACTUAR J
[4]
BENJAMIN B., 1995, MORTALITY MOVE
[5]
Benjamin B., 1980, ANAL MORTALITY OTHER
[6]
BUUS H, 1960, T 25 INT C ACT, V2, P364
[7]
Copas J., 1983, J. Inst. Actuaries, V110, P135
[8]
CRAMER H, 1935, SKAND AKTUARIETIDSK, P161
[9]
FELIPE A, 1998, T 26 INT C ACT, V9, P250
[10]
Forfar D., 1988, T FACULTY ACTUARIES, V40, P98