1
$\begingroup$

I'm using the MissRanger package in R to perform imputation on a very large dataset. While the package documentation clearly explains the iterative random forest imputation process, I couldn't find explicit details regarding how MissRanger initializes missing values before starting the iterations.

My intuition, based on reading the documentation, suggests that MissRanger might be using the imputeUnivariate() method as this is part of the MissRanger package. imputeUnivariate() is described as:

Fills missing values of a vector, matrix or data frame by sampling with replacement from the non-missing values. For data frames, this sampling is done within column.

Could someone clarify how MissRanger initializes the imputation process?

Improve this question
$\endgroup$
3
  • $\begingroup$ Welcome! Did you try asking this on Stack Overflow? Since it's mainly programming based, you may find a better answer over there. $\endgroup$ Commented Apr 4 at 14:23
  • $\begingroup$ Hi Rick. Yes, I did, but it was deleted with the message: "Closed. This question is not about programming or software development. It is not currently accepting answers." I've been reading posts from Stack Overflow and Stack Exchange for years, as they always appear on Google, but I've never asked any questions before—so I'm kinda unsure how to use it myself. $\endgroup$ Commented Apr 9 at 9:23
  • $\begingroup$ No worries. This kind of post is always hard because it’s like right in the middle! ;) $\endgroup$ Commented Apr 9 at 14:02

1 Answer 1

Reset to default
3
$\begingroup$

Judging from the code, it uses imputeUnivariate for the first variable then imputes the second variable using the first, the third using the first two, and so on until there are initialisations for everything.

Improve this answer
$\endgroup$
2
  • $\begingroup$ I can confirm that this is correct. $\endgroup$ Commented Apr 5 at 10:58
  • $\begingroup$ Thank you Thomas and Michael for the confirmation! $\endgroup$ Commented Apr 9 at 9:41

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.