本文在一些弱的条件下,对自然联系函数和自适应设计下广义线性模型的极大拟似然估计渐近性进行研究,获得了极大拟似然估计的渐近存在性、弱相合性、收敛速度及渐近正态性.并通过蒙特卡罗数值模拟的方法对所得结果进行验证.
In this paper, for the generalized linear models(GLMs) with natural link and adaptive designs, we consider the asymptotic properties of maximum quasi-likelihood estima- tors(MQLEs) under some mild conditions. The asymptotic existence of MQLEs in quasi- likelihood equation, weak consistency, the rate of convergence and asymptotic normality of MQLEs are presented. The results are illustrated by Monte-Carlo simulations.