The start of a Julia package for the Multiple Imputation by Chained Equations algorithm.

Language: Julia
Source code: GitHub

The start of an implementation of the MICE algorithm in Julia.


The plan for this package is to compare its output to other iterative imputation packages, including the R package mice and scikitlearn's IterativeImputer.

To generate reproducible comparison datasets, the mice package is run in a docker container.

cd test/comparison/
# Build the container
docker build -t mice_jl_r_script .
# Run the container
docker run --rm -v "$(pwd)/data/:/data/" mice_jl_r_script


If you want to learn more about the MICE algorithm, these are the main sources I am using:

  • Azur, M. et al. (2011). Multiple imputation by chained equations: what is it and how does it work?. Link
  • van Buuren, S. (2018). Flexible Imputation of Missing Data, Second Edition (2nd ed.). Link