Model averaging for generalized linear models in fragmentary data prediction

Fragmentary data is becoming more and popular in many areas which brings big challenges to researchers analysts. Most existing methods dealing with fragmentary consider a continuous response while applications the variable discrete. In this paper, we propose model averaging method for generalized linear models prediction. The candidate are fitted based on different combinations of covariate availability sample size. optimal weight selected by minimizing Kullback–Leibler loss completed cases its asymptotic optimality established. Empirical evidences from simulation study real analysis about Alzheimer disease presented.