摘要
本文在响应变量的观测值为Ⅰ型区间删失数据的情形下,讨论部分线性模型Sieve极大似然估计的渐近性质.在一定条件下证明了该估计具有强相合性;参数分量的估计具有渐近正态性,并且是渐近有效的;非参数分量估计达到了最优弱收敛速度.
For a partly linear model, when observations of the respond variable are case one interval censored, asymptotic properties of sieve MLE are discussed. Piecewise linear functions are used to struct sieve spaces, under mild conditions, sieve MLE are shown to be strong consistent, sieve MLE of the nonparametric part has an optimal convergence rate, sieve MLE of the parametric part is asymptotic efficient.
出处
《应用数学学报》
CSCD
北大核心
2001年第1期139-151,共13页
Acta Mathematicae Applicatae Sinica
基金
国家自然科学基金资助项目.