IMPUTACIÓN MÚLTIPLE EN VARIABLES CATEGÓRICAS USANDO DATA AUGMENTATION Y ÁRBOLES DE CLASIFICACIÓN

Abstract

It is a modication of common multiple imputation algorithms which combines classification trees (CT) and data augmentation for categorical data. We describe the rationale of the method and compare it, on theoretical and practical grounds, with two of the most frequently used methods. We use a fictitious base and an “ad hoc" R-based software