Investigating the Performance of a Variation of Multiple Correspondence Analysis for Multiple Imputation in Categorical Data Sets

Non-response in survey data, especially in multivariate categorical variables, is a common problem which often leads to invalid inferences and inefficient estimates. A regularized iterative multiple correspondence analysis (RIMCA) algorithm in single imputation (SI) has been suggested for the handli...

Full description

Saved in:
Bibliographic Details
Published in:Journal of classification 2017-10, Vol.34 (3), p.384-398
Main Authors: Nienkemper-Swanepoel, Johané, von Maltitz, Michael J
Format: Article
Language:English
Subjects:
Algorithms
Bioinformatics
Computer simulation
Datasets
Marketing
Mathematics and Statistics
Maximum likelihood method
Multivariate analysis
Pattern Recognition
Psychometrics
Regression analysis
Signal,Image and Speech Processing
Statistical Theory and Methods
Statistics
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Non-response in survey data, especially in multivariate categorical variables, is a common problem which often leads to invalid inferences and inefficient estimates. A regularized iterative multiple correspondence analysis (RIMCA) algorithm in single imputation (SI) has been suggested for the handling of missing categorical data in survey analysis. This paper proposes an adapted version of the SI algorithm for multiple imputation (MI). The SI and MI techniques are compared for both simulated and real questionnaire data. A comparison between RIMCA MI and Sequential Regression Multiple Imputation (SRMI) is shown to establish the success of the proposed MI procedure.
ISSN:0176-4268
1432-1343
DOI:10.1007/s00357-017-9238-6