Abstract:
|
Multiple imputation (MI) is a popular method for dealing with missing data in sample surveys. MI by chained equations (MICE) is one of the most widely used MI algorithms for multivariate data, but it lacks theoretical foundation and is computationally intensive. Recently, missing data imputation methods based on deep learning models have been developed with encouraging results in small studies. However, there has been limited research on systematically evaluating their performance in realistic settings compared to MICE, particularly in large-scale surveys. We conduct extensive simulation studies to compare the repeated sampling properties of four machine learning based MI methods: MICE with classification trees, MICE with random forests, generative adversarial imputation network, and MI using denoising autoencoders. We find the deep learning based methods dominate MICE in terms of computational time; however, under the default choice of hyperparameters in the common software packages, MICE with classification trees consistently outperforms the deep learning imputation methods in terms of bias, mean squared error, and coverage under a range of realistic settings.
|