synthpop (version 1.9-0)

mergelevels.syn: Merge levels of factors in a data frame

Description

merges levels of selected variables in a data frame either according to minimum numbers in a category or according to user-defined rules.

Usage

mergelevels.syn(data, vars = NULL, newlabel = FALSE, addNA = FALSE,
                    print.flag = FALSE, minsize = 10, merge.byhand =
                    FALSE, merge.details = NULL)

Value

A data frame of the same size and structure as data with levels of selected variables merged.

Arguments

data

An observed data set before synthesis.

vars

a vector of names or numbers for the variables for which categories are to be merged. defaults to all factors in data when sert to NULL

minsize

The minimum size that of combined categories when merge.byhand is FALSE

newlabel

When merge.byhand is FALSE when newlabel = TRUE causes the new factor levels to be labelled with a concatenation of the labels of all the categories that have contributed. newlabel = FALSE labels the new category "Other"

addNA

Causes the NA category to be included when determining which groups are below minsize and also to allow NA values, designated by "NA", to be combined with others in merge.details

print.flag

prints tables of variables before and after recoding.

merge.byhand

Uses the information in merge.byhand to create groupings, ignoring the parameters minsize and vars

.

merge.details

A named list of variable names with names giving the names of the variables that will have levels merged. Each item is a vector where the first item is the name for the new combined level and the other entries are the levels to be merged. If it exists already the levels with small counts will be added to it, otehrwise a new level will be formed.

See Also

sdc

Examples

test <- SD2011[1:20]
data.mlevs1 <- mergelevels.syn(test, vars = c(3,5,18:20),minsize = 20, addNA = TRUE, 
              print.flag = TRUE, newlabel = TRUE)

mlevs <- list(agegr = c("60+", "60-64"  , "65+"),socprof = c("NEW","UNEMPLOYED","FARMER"))
data.mlevs2 <-mergelevels.syn(test, merge.byhand = TRUE, merge.details = mlevs, 
           addNA=TRUE, print.flag = TRUE)