Fractional Imputation for Multivariate Categorical Missing Data

Tuesday, January 28, 2020 - 12:30pm to 1:00pm
Event Type: 

Speaker: Lei Zhou

Abstract: Multivariate categorical missing data are prevalent in practice. In this talk, we briefly review the related concepts in categorical missing data and imputation methods. We will illustrate the joint modeling approach and conditional modeling approach and show why conditional modeling is computationally efficient and tractable under certain scenarios in high dimensional settings. Our simulation result shows the proposed method performs better than multiple imputation using the Dirichlet process model.