Estimation under length-bias and right-censoring:: an application to unemployment duration analysis for married women
Estimation under length-bias and right-censoring:: an application to unemployment duration analysis for married women
复制标题
DOI：
10.1080/0266476022000030066
复制
发表时间：
2003-04-01
影响因子：
1.5
通讯作者：
Alvarez-Llorente, G
中科院分区：
文献类型：
--
作者：
de Uña-Alvarez, J;Otero-Giráldez, MS;Alvarez-Llorente, G
在这项工作中，我们分析了西班牙已婚妇女的失业时间，使用劳动力调查（LFS）的西班牙统计研究所，1987-1997年。一致的非参数估计的失业生存函数。由于可用数据存在长度偏倚，因此对Kaplan-Meier乘积限估计量进行适当校正，并用于参考分析。参数模型的准确性通过拟合优度图来检查，这是一种需要对生存率进行初步估计的图形工具。探索了相关危险的结构特征（如单调性和单峰性）。
In this work we analyse unemployment duration for married women in Spain, using the Labour Force Survey (LFS) of the Spanish Institute for Statistics, 1987-1997. Consistent non-parametric estimation of the unemployment survival function is provided. Since the available data are length-biased, a suitable correction of the Kaplan-Meier product-limit estimator is motivated and used for the referred analysis. The accuracy of parametric models is checked by means of goodness-of-fit plots-a graphical tool that requires preliminary estimation of the survival. Structural features of the associated hazard (as monotonicity and unimodality) are explored.