resample.data.frame {MIfuns}R Documentation

Create Replicate Data Sets by Stratified Sampling With Replacement

Description

resample() is generic. A Methods is defined for data.frame; a convienience wrapper is provided for passing names of csv files to be read and then resampled.

Usage

        ## S3 method for class 'character':
        as.csv.filename(x, ...)
        ## S3 method for class 'csv.filename':
        resample(x, ...)
        ## S3 method for class 'data.frame':
        resample(x, names, key = NULL, rekey =
      FALSE, out = NULL, stratify = NULL, seed = 0, ...)

Arguments

x A data.frame, or (second form) a file name for a CSV file to read.
names A list of names for replicate data sets; can be a simple vector.
key An atomic character value naming the column in x that distinguishes unique individuals, (resampling targets). Defaults to row names.
rekey If true, key values in resampled data sets will have unique values of key replaced with consecutive integers, starting at 1.
out A (path and) directory in which to write resulting data sets as CSV files.
stratify A list of factors, the interactions of which will be the levels of stratification. Each factor must have the same length as nrow(x). Or a character vector of names in names(x).
seed A number to use as the seed for the random number generator.
... Extra arguments, currently ignored.

Details

Typical usages are

                resample(x, names, ...)
                resample(as.csv.filename(x), names, ...)
The argument key gives the name of the column in x to identify unique experimental units (individuals). If not supplied, a temporary key is constructed from the row names, and sampling occurs at the row level.

The number of resamplings is controlled by the length of names. names is coerced to character, and each value is used to name a *.csv file, if out is supplied. If out is omitted, a list of data.frame is returned.

stratify is a list of factors, or items that can be coerced to factors. Currently stratify is coerced to a data.frame for convenient manipulation. Empty levels are dropped. If stratify is not supplied, the whole data set is treated as a single level. Otherwise, each resulting data set has as many keys in each level as the original. An error results if key is not nested within stratify.

Value

A list of data.frame, or if out is supplied, an invisible list of the numbers of rows of each data.frame written to file.

Note

resample.character, etc. has been moved to metaSub.character, etc.

Author(s)

Tim Bergsma

References

This package developed under the auspices of Metrum Institute, http://metruminstitute.org

See Also

metaSub.character

Examples

b <- resample(Theoph,key="Subject",names=1:3)
d <- resample(
        Theoph,
        key="Subject",
        rekey=TRUE,
        names=1:3,
        out=".",
        stratify=Theoph$Dose < mean(Theoph$Dose)
)
e <- resample(as.csv.filename("1.csv"),names="theoph")

[Package MIfuns version 3.2.9 Index]