Nonparametric Iterative Imputation Methods for Missing data

This article was an attempt to describe nonparametric iterative imputation methods and compare them to complete-case analysis and prevalent imputation methods for missing data. The data of 478 households in the cost and income sampling of rural and urban household in Shahrekord in the year 1395 were used for research. Data sets had single structured model, but varying in size of missing value. These methods were then compared in terms of their ability to reconstruct the original data. The results were significant while using nonparametric iterative imputation methods. After first iteration, MSE, Bias and Variance of parameter decreased so fast, and they kept decreasing until approximate converges. Furthermore, all these nonparametric iterative imputation methods led to same estimators and same estimates of missing values.

Keywords:

Missing data , Nonparametric Iterative Imputation , Complete-Case analysis , MSE , Bias , Variance