Investigating Some Imputation Methods of Multivariate Imputation Chained Equations

This paper investigates three MICE methods: Predictive Mean Matching (PMM), Quantile Regression-based Multiple Imputation (QR-basedMI) and Simple Random Sampling (SRSI) at imputation numbers 5, 15, 20 30 with 5% 20% missing values, to ascertain the one that produces imputed values best matches observed compare model fit based on AIC MSE. The results show that; QR-basedMI produced more didn’t match observed, SRSI better as number of imputations increases while PMM matched all missingness considered. for showed in terms MSE except M=15, result M= M=15 M=20 results. shows considered both where was seen produce It is concluded comparison, most suited when but QR-basedMI.