广义估计方程根的强相合性

Strong Consistency of Generalized Estimation Equation Root

在广义估计方程有正确假定,当个体数目趋于无穷而残差为鞅差序列时,用鞅伯恩斯坦不等式证明广义估计方程根的强相合性,同时在个体数目为1而重复观测次数趋于无穷且残差为鞅差序列时,也证明了广义估计方程根的强相合性.

The paper studies strong consistency of generalized estimation equation root.Under the generalized estimating equation having correct assumption,when the number of subjects goes to infinity and the residuals form a martingale difference sequence,the paper uses martingale Benstein Inequality to prove strong consistency of the generalized estimation equation root.At the same time when the number of subjects is one,and the number of observations on the subject goes to infinity,and the residuals form a martingale difference sequence,the paper proves strong consistency of generalized estimation equation root.