摘要
本文基于1981—2010年台湾人口死亡率数据,对人口死亡率估计模型进行深层次分析。研究发现:⑴1-100岁之间各年龄组别中心死亡率与时间因素存在显著负相关;45岁以后,时间对各年龄组别中心死亡率的影响程度形成"V"字型。⑵不同年龄组和不同时间段之间过渡时所呈现的"亦此亦彼"性,即模糊性,对死亡率的影响在现有的死亡率估计模型中常被忽视。⑶死亡率的不确定性主要源自模糊不确定性,本文构建的死亡率模糊模型可以提高死亡率估计精度。
The paper makes a detailed study on the mortality models based on Taiwan population mortality from 1981 to 2010. The findings show that there has been a significantly negative correlation between time factor and central mortality in dif- ferent age groups from 1 to 100 years old. Time influence degree on central mortality in different age groups over 45 years old exercises like a V shape. On the other hand, there appears "fuzziness" that happens at a transitional period of time between different age groups and different time phases, whose influence on death rate tends to be ignored in the existing mortality mod- els. Finally, the uncertainty of mortality comes mainly from fuzziness, and fuzzy mortality model produced in this paper can im- prove the estimation accuracy of mortality.
出处
《台湾研究集刊》
CSSCI
2014年第3期87-94,共8页
Taiwan Research Journal
基金
国家社科基金重大项目"大数据与统计学理论的发展研究"(13&ZD148)
"计量经济学"教育部重点实验室(厦门大学)
福建省统计学重点实验室(厦门大学)
关键词
中心死亡率
模糊不确定性
反模糊化
central mortality, fuzzy uncertainty, defuzzification