PcAux

Archive Notice

PcAux development no longer uses this repository. So, the repository has been archived. You can find the new home of PcAux development at https://github.com/Statscamp/PcAux.

Below, you can find the original README information for this repository.

This is the repository for the PcAux package which was formerly called "quark."

• Licensing information is given in the LICENSE file.
• Built tarballs of the PcAux package are available in the builds directory.
• Stand-alone documentation is available in the documentation directory.
• The source files for the most recent stable version of PcAux are available in the source directory.

PcAux is beta software, so please report any bugs that you encounter in the issues section of the project page. You may also leave requests for new features in the issues section.

Thank you for your interest in the PcAux project! I hope you find our software useful!

Installation

The best way to install PcAux is to use the devtools::install_github function.

1. First, make sure that you have devtools installed on your system

2. Next, execute the following lines:

   library(devtools)
   install_github("PcAux-Package/PcAux/source/PcAux")

3. Finally, load PcAux and enjoy:

   library(PcAux)

If the devtools-based approach does not work, you can download one of the built tar-balls from the builds directory and manually install the package from source by executing the following lines:

    install.packages("/SAVE_PATH/PcAux_VERSION.tar.gz",
                     repos = NULL,
                     type  = "source")

Where SAVE_PATH is replaced by the (relative or absolute) file path to the location where you saved the tar-ball, and VERSION is replaced with the correct version number for the tar-ball that you downloaded.

Example

A basic missing data treatment using PcAux might look like the following:

1. First, load and prepare your data:

   data(iris2)
   cleanData <- prepData(rawData   = iris2,
                         nomVars   = "Species",
                         ordVars   = "Petal.Width",
                         idVars    = "ID",
                         dropVars  = "Junk",
                         groupVars = "Species")

2. Next, create a set of principal component auxiliary variables:

   pcAuxOut <- createPcAux(pcAuxData = cleanData,
                           nComps    = c(3, 2))

3. Finally, use the principal component auxiliaries as the predictors in a multiple imputation run:

   miOut <- miWithPcAux(rawData   = iris2,
                        pcAuxData = pcAuxOut,
                        nImps     = 5)

You can also work directly with the principal component auxiliaries:

• You can merge the principal component auxiliaries back onto your raw data (e.g., for use with the Graham, 2003, saturated correlates approach).

  outData <- mergePcAux(pcAuxData = pcAuxOut, rawData = iris2)

• You can also create a stand-alone predictor matrix that can be used to correctly incorporate the principal component auxiliaries into a separate MI run using the mice package.

  predMat <- makePredMatrix(mergedData = outData)