摘要
本文在不完全数据的情况下采用Ｌ１模得到了回归函数的Ｐａｒｔｉｔｉｏｎｉｎｇ估计的强相合性．我们的结果只要求回归函数的Ｐ（Ｐ＞１）阶矩有限，对回归函数没加任何光滑条件，并且我们的结果对任何分布都成立．
The strong consistencies in L1-Norm of Partitioning estimate of a reqression unction are shown for incomplete sample. Here incomplete sample means that some samplepoints are missing according to a random missing mechanism operated on the observationvectors, and our results only ask the finity of pth-moment (for some p>1) and some conditions on random missing mechanism operator, do not ask any smooth condition on regressionfunction and any condition on distribution.
出处
《黑龙江大学自然科学学报》
CAS
1996年第1期10-15,共6页
Journal of Natural Science of Heilongjiang University
关键词
强相合
回归函数
随机损失数据
P估计
Strong consistency
Partitioning Estimate
Reqression Function
RandomiyMissing Date
