I was trying to use 'importance' parameter for imputation with the miceRanger package. However, in the runIterations function, the importance parameter is hardcoded and doesn't respond to changes. Specifically, within the rangerArgs list, the importance argument is set to "impurity" like this:
rangerArgs <- list(
data = dats[!missIndx, algCols, with = FALSE],
dependent.variable.name = impVar,
importance = "impurity",
probability = returnProb,
verbose = FALSE,
...,
max.depth = 10
)
As a result, any attempt to modify the importance setting in the miceRanger() function has no effect.
In my opinion the 'importance' parameter in rangerArgs should be set to importance - the same which is set in the function miceRanger
I was trying to use 'importance' parameter for imputation with the miceRanger package. However, in the runIterations function, the importance parameter is hardcoded and doesn't respond to changes. Specifically, within the rangerArgs list, the importance argument is set to "impurity" like this:
rangerArgs <- list(
data = dats[!missIndx, algCols, with = FALSE],
dependent.variable.name = impVar,
importance = "impurity",
probability = returnProb,
verbose = FALSE,
...,
max.depth = 10
)
As a result, any attempt to modify the importance setting in the miceRanger() function has no effect.
In my opinion the 'importance' parameter in rangerArgs should be set to importance - the same which is set in the function miceRanger