摘要
研究了在删失样本下误差为鞅差序列时 ,回归函数加权核估计的r阶矩收敛性 ,完全收敛性和几乎处处收敛性 ,推广了在完全样本下误差为鞅差序列时相应的结论 ,同时还给出了r(r >1)
The r order mean consistency, the complete convergence and strong consistency of the weighted kemel estimator for regression function, under martingale difference error sequences from censored data are studied. These conclusions generalize the results under the complete sample. The r order mean consistency rate of this estimator is also obtained.
出处
《宁波大学学报（理工版）》
CAS
2002年第2期4-6,共3页
Journal of Ningbo University：Natural Science and Engineering Edition
关键词
回归函数
加权核估计
删失样本
鞅差序列
r阶矩收敛性
完全收敛性
几乎处处收敛性
weighted kernel estimator
randomly censored data
martingale difference sequences
r order mean consistency
complete convergence
strong consistency
