非参数回归函数估计的随机加权逼近

Random Weighting Approximation for the Estimates of Nonparamatric Regression Function

本文应用随机加权法的思想构造了回归函数的最近邻估计和核估计的随机加权统计量，并证明了用随机加权统计量的分布逼近两类估计量的分布之精度可达到ｏ（ｎ^（－１／２Ｐ），其中１＜ｐ ≤ ２．

Applying the random weighting methods we structure random weighting statistic of the esti- mates of repression function, and prove that the random weighting approximation can attain the accuracy of o(n^(-1/2p)), where 1 < p ≤2.