Title: Model Averaging for Generalized Linear Model with Covariates that are Missing completely at Random

Abstract: In this paper, we consider the estimation of generalized linear models with covariates that are missing completely at random. We propose a model averaging estimation method and prove that the corresponding model averaging estimator is asymptotically optimal under certain assumptions. Simulaiton results illustrate that this method has better performance than other alternatives under most situations.