Write a List of Multiply Imputed Datasets
Writes a list of multiply imputed datasets.
write.datlist(datlist, name, include.varnames=TRUE, type="csv2",
separate=TRUE, Mplus=FALSE, round=NULL, Rdata=TRUE,
subdir=TRUE, ...)datlist |
List of imputed datasets. Can also be an object of class
|
name |
Name of files to be saved |
include.varnames |
Logical indicating whether variables should be saved |
type |
File type of datasets to be saved, see |
separate |
Logical indicating whether imputed datasets should be written in separate files. |
Mplus |
Logical indicating whether files should be written for usage in Mplus software |
round |
Number of digits to round after decimal. The default is no rounding. |
Rdata |
Logical indicating whether |
subdir |
Logical indicating whether results should be written into a subdirectory. |
... |
Further arguments to be passed to |
See also mice::mids2mplus,
mice::mids2spss and
write.mice.imputation for writing objects of class
mids.
See also Amelia::write.amelia for
writing imputed datasets in Amelia.
## Not run:
#############################################################################
# EXAMPLE 1: Write data list imputed in mice
#############################################################################
data(data.ma01)
dat <- as.matrix(data.ma01)
# start with empty imputation
imp0 <- mice::mice( dat, maxit=0)
# modify predictor matrix
predM <- imp0$predictorMatrix
predM[, c("idschool", "idstud" ) ] <- 0
# modify imputation method
impMeth <- imp0$method
impMeth[ impMeth=="pmm" ] <- "norm"
# do imputations in mice
imp <- mice::mice( dat, predictorMatrix=predM, method=impMeth, m=3, maxit=4 )
# write imputed data in format "csv2" and round after 4 digits
write.datlist( datlist=imp, name="mice_imp_csv2", round=4 )
# write imputed data in R binary format
write.datlist( datlist=imp, name="mice_imp_Rdata", type="Rdata")
# write data for Mplus usage
write.datlist( datlist=imp, name="mice_imp_Mplus", Mplus=TRUE, round=5)
## End(Not run)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.