最大似然估计迭代公式的收敛性及其应用

Convergence of Iteration Methods of Maximum Likelihood Estimator and Its Applications

本文研究了最大似然估计量常用迭代公式的收敛性问题.给出了Gauss-Newton迭代公式具有渐近数值稳定性的充分条件.在此基础上,研究了改进的Gauss-Newton迭代公式及其收敛的充分条件.通过实例计算了Weibull分布以及广义线性模型的最大似然估计,取得很好的效果.

Iteration methods and their convergences of the maximum likelihood estimator are discussed in this paper. We study the Gauss-Newton method and give a set of sufficient conditions for the convergence of asymptotic numerical stability. The modified Gauss-Newton method is also studied and the sufficient conditions of the convergence are also presented. Two numerical examples are given to illustrate our results.