Abstract: This article studies the consistency, asymptotic normality. hypothesis testing of M-estimate of themultivariate linear resression coefficient by minimizing the randomly weighted sum (Yi-X’iβ) . and the Bootstrap approximation to the distribution of nonrandom weighted M-estimate.

Key words: multivariate linear niodel, M-estimate. randomly weighted method . asymptotic normality

摘要： 本文讨论了由随机加权最小问题的解定义的多元线性回归系数的随机加权Ｍ－估计的相合性、渐近正态性，参数检验及时非随机加权Ｍ－估计分布的Bootstrap逼近。

关键词: 多元线性模型, M－估计, 随机加权, 渐近正态性