'Should I impute missing values and then log transform the variables, or the other way around
I'd like to train three different models:
- elastic net regression, caret, method = 'glmnet'
- Random Forest, caret, method = 'ranger'
- Simple linear regression
on a dataset containing relatively high number of missing values.
I found the package missRanger to handle the missing values.
However, I am not sure whether I should log-transform the data first and then impute the NAs, or the other way around?
Sources
This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.
Source: Stack Overflow
| Solution | Source |
|---|