摘要
本文考虑了标准化的多参数Ｍ估计之误差分布的随机加权逼近，得到的逼近精度为ｏ（ｎ￣（－１／２）），从而我们推广了［１］和［２］的结果。
In this paper we have considered the random weighting approximation for the error distribution of normalized multiparametric M-estimator,the approxi-mate accuracy has been proved to be at the optimal rate of o(n￣(-/1)).We have not on-ly extended[1]and[2]’s results to more general cases but also given an easier way to construct the random weighting statistic without knowing the Edgeworth expan-sion.
出处
《北京大学学报（自然科学版）》
CAS
CSCD
北大核心
1994年第5期509-518,共10页
Acta Scientiarum Naturalium Universitatis Pekinensis
基金
国家自然科学基金
关键词
M估计
随机加权
参数估计
M-estimation
random weighting
Edgeworth expansion
