Model Averaging for Generalized Linear Models in Fragmentary Data Prediction

TL;DR

This paper introduces a model averaging approach for generalized linear models tailored to fragmentary data, effectively handling discrete responses and varying covariate availability, with proven asymptotic optimality.

Contribution

It proposes a novel model averaging method for GLMs in fragmentary data scenarios, addressing discrete responses and establishing asymptotic optimality.

Findings

Method performs well in simulations

Effective on Alzheimer disease data

Achieves asymptotic optimality

Abstract

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