setRFoptions {CALIBERrfimpute} | R Documentation |
Set Random Forest options for imputation using MICE
Description
A convenience function to set global options for number of trees or number of nodes.
Usage
setRFoptions(ntree_cat = NULL, ntree_cont = NULL,
nodesize_cat = NULL, nodesize_cont = NULL,
maxnodes_cat = NULL, maxnodes_cont = NULL)
Arguments
ntree_cat |
number of trees to be used for imputing categorical variables (each imputed value is the prediction of a randomly chosen tree), default = 10. |
ntree_cont |
number of trees in the forest for imputing continuous variables, default = 10. |
nodesize_cat |
minimum node size for trees for imputing categorical variables, default = 1. A higher value can be used on larger datasets in order to save time. |
nodesize_cont |
minimum node size for trees for imputing continuous variables, default = 5. A higher value can be used on larger datasets in order to save time. |
maxnodes_cat |
maximum number of nodes in trees for imputing categorical variables. By default the size limit is set by the number of observations and nodesize_cat. |
maxnodes_cont |
maximum number of nodes in trees for imputing continuous variables. By default the size limit is set by the number of observations and nodesize_cont. |
Details
This function sets the global options which have the prefix 'CALIBERrfimpute_'.
Value
No return value. The function prints a message stating the new option setting.
See Also
mice.impute.rfcat
, mice.impute.rfcont
Examples
# Set option using setRFoptions
setRFoptions(ntree_cat=15)
options()$CALIBERrfimpute_ntree_cat
# Set option directly
options(CALIBERrfimpute_ntree_cat=20)
options()$CALIBERrfimpute_ntree_cat