3

I've found the package Impute.jl but it's only able to use these simple methods:

  • drop: remove missing.
  • locf: last observation carried forward
  • nocb: next observation carried backward
  • interp: linear interpolation of values in vector
  • fill: replace with a specific value or a function...

There seems not to exist any advanced "multiple imputation" method.
How can use more advanced methods when I have several variables?
Such as: fully conditional specification (mice), bayesian methods, random forest, multilevel, nested imputation, censored data, categorical data, survival data...

I don't mean creating my own code but finding any Julia package able to do it automatically. Other software do have it (R, Python, Stata, SAS…).

skan
  • 7,423
  • 14
  • 59
  • 96

1 Answers1

1

See the Imputation module of BetaML.

It includes imputation for tabular data using Gaussian Mixture Models, Random forests or any other ML model with a fit!/predict interface.

It includes imputations of multiple columns, optionally with multiple imputations and multiple "passages" trough the dataset to impute.

(disclaimer: I'm the author...)

Antonello
  • 6,092
  • 3
  • 31
  • 56