Random forest modeling with the ranger package does not allow for missing values. This issue can be solved using the missRanger function (see rangerForest.R). However, because missRanger uses ensemble models just to impute the missing values, it takes a huge amount of time to run even with some optimization tweaks.
[ ] Decide how we want to deal with missing values
Random forest modeling with the ranger package does not allow for missing values. This issue can be solved using the missRanger function (see rangerForest.R). However, because missRanger uses ensemble models just to impute the missing values, it takes a huge amount of time to run even with some optimization tweaks.