misle (Depreciated)

Multiple imputation through statistical learning

misle has been depreciated and it is now separated into the following packages

1. mixgb: multiple imputation through XGboost
2. miae: multiple imputation through auto-encoders including
   • midae: multiple imputation by denoising autoencoders(with dropout)
   • mivae: multiple imputation by variational autoencoders
3. vismi: Visualsation Tools for Multiple Imputation

mixgb

multiple imputation through XGboost

Mixgb is now on CRAN and the newest version is on Github. For more details please check https://github.com/agnesdeng/mixgb

miae

multiple imputation through autoencoders

Under development :) https://github.com/agnesdeng/miae

• Using torch R package (pytorch) instead of tensorflow
• Writing up documentation and vignette
• Pretune hyperparamters for imputers
• Visual diagnostic for imputation results

vismi

Visualsation Tools for Multiple Imputation

https://github.com/agnesdeng/vismi