Simulation Study on Missing Data Imputation for Dichotomous Variables using Statistical and Machine Learning Methods in Scientific Reports

A new study has investigated the use of statistical and machine learning methods for imputing missing data in dichotomous variables. The study, published in Scientific Reports, used simulated data to generate a variety of missing scenarios and compared the performance of three traditional statistical methods and five machine learning methods for data imputation. The study considered multiple factors including missing mechanisms, sample sizes, missing rates, the correlation between variables, value distributions, and the number of missing variables. The results of the study shed light on the imputation performance of the different methods and could be useful for researchers in a range of fields.