Advanced Missing Value Imputation Techniques: Machine Learning Methods with an Emphasis on an Ensemble Method for Multiple Imputation by Chained Equations

IntroutionMissing data is a well-known problem common in all organizations that gather data on persons or enterprises.It causes issues like performance degradation, analysis challenges, issues with data processing, and biased results due to the disparities between missing and complete values. Material and MethodsTraditional imputation techniques, including mean, regression, K-nearest neighbor, and ensemble methods, have been proposed to handle missing values.Multiple imputation by chained equations (MICE) is one of the most common methods for imputation.In theory, any imputation model can be used to predict the missing values.However, if the predictive models are incorrect, it can lead to biased estimates and invalid inferences.One of the latest solutions for dealing with missing data is machine learning methods and the SuperMICE method, which is a combination of the two methods of SuperLearner and MICE. Results and DiscussionIn the random mechanism, when missingness is 10%, 20%, and 30%, the performance of PMM, SuperMICE, and BLR is similar.However, at higher missingness levels (40% and 50%), SuperMICE performs better, with a significant difference compared to the other two methods.In real data with a . . . . . . . . .