自适应广义线性回归拟似然估计渐近理论

Asymptotic Theory of Quasi-maximum Likelihood Estimator in Generalized Linear Models With Adaptive Designs

研究自适应设计下的拟似然方程 ∑ni=1xi(yi- μ(x′iβ) ) =0 ,在一定的条件下证明了以概率 1此方程当n充分大时有解^βn,^βn 为β0 的强相合估计 ,且得出了 ^βn- β0的收敛速度 ;然后又在一定的条件下证明了 ^βn

This paper studies the quasi likelihood equation ∑ni=1x i(y i-μ(x ′ iβ))=0 with adaptive designs.Under certain conditions, we obtain that there exists solution n to the above equation with probability 1 for sufficently large n, n is the estimator of strong consistency of β 0, and the convergence rate of n-β 0; And then ,under certain conditions we obtain some results of asymptotic normality.