广义线性模型参数估计的随机加权法

Randomly weighted method in parameter estimation of generalized linear model

在广义线性模型中,用随机加权法求解参数估计的渐近方差.通过推导给出随机加权参数估计与未加权参数估计的渐近分布,并且证明它们是相同的,从而可以用容易获得的随机加权估计的渐近分布去近似原参数估计的渐近分布.

In the generalized linear model , the randomly weighted method was applied to solve the asymptotic variance of parameter estimation .The asymptotic distribution of random weighted parameter estimation and unweighted parameter estimation were given by derivation ,and they were proven to be the same .So the easily acquired asymptotic distribution of the random weighted estimation could be used to approximate the asymptotic distribution of the original parameter estimation .