Distribution of multidimensional data #621
enricorubagotti
started this conversation in
Missing data methodology
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I am planning to use the documented mice 2.9 to impute a 50K plus dataset with >100 fields mixed as continuous (>0) and nominal with multiple levels.
I do not believe the continuous data to be distributed as a gaussian. I would like a clarification on the sentence "Note that it is possible to specify models for which no known joint distribution exits." quoted from the article "mice: Multivariate Imputation by Chained Equations in R". Is the gaussian distribution a requirement to use mice?
Beta Was this translation helpful? Give feedback.
All reactions