The plugin adds a Design menu to the R-Commander. From this, various types of experimental plans can be generated, imported, augmented with responses or further runs, and - in a limited way - analysed. Of course, other R-Commander functionality can also be used on the designs.
The plugin is currently a beta version. Reports on bugs or inconveniences are very welcome.
The R GUI should be installed with the Single Document Interface (SDI) mode instead of the default MDI mode. John Fox, the author of R-Commander, has described how to change MDI to SDI, if R has been installed in the wrong mode. It is strongly discouraged to run the R-Commander under the MDI mode, as it happens very frequently that windows suddenly disappear. They usually have iconized only and can be retrieved from the taskbar; nevertheless, this behavior can be very annoying and can be avoided by using R in SDI mode.
For ensuring valid checks of package RcmdrPlugin.DoE, it was necessary to include package FrF2.catlg128 under suggested packages. Note that most users will not need that package and should not install it.
RcmdrPlugin.DoE is currently in beta shape. The calculation functions have been reasonably tested within the underlying calculation packages. Usage of the menus has of course also been tested, but there are very likely untested scenarios where something does not work properly. Bug reports are therefore invited!
The logged commands are currently aiming to be correct but not to be didactic,
since they have been generated with the purpose to obtain a valid result from the inputs with as
little programming effort as possible. They often contain many superfluous inputs,
since default inputs could also be omitted. For example, while FrF2(8,4) would
produce a regular fractional factorial design with 8 runs, 4 factors and default factor names and levels,
the respective logged command is three rows long.
The help files will have to be improved upon. Suggestions are welcome (where did you get stuck?).
Current users of this package have to accept acting as beta testers. It is particularly important to never work with the one and only data original, but have the valuable experimental data themselves stored in a safe place.
Full factorial designs, orthogonal main effects designs, regular and non-regular 2-level fractional factorial designs, central composite and Box-Behnken designs, latin hypercube samples, and simple D-optimal designs can currently be generated from the GUI.
Extensions to cover further latin hypercube designs as well as more advanced D-optimal designs (with blocking) are planned for the future.
RcmdrPlugin.DoE implements functionality from packages DoE.base, FrF2, DoE.wrapper (which in turn implements functionality from several other packages), and also from package rsm. All package allow more flexibility through command line programming than through using the GUI. Nevertheless, usage of the GUI should be sufficient for reasonable creation and analysis of results at least in standard situations.
There is a tutorial (see references) with example data; it is, however, far from being comprehensive. For experienced users of other DoE Software, the menus may be reasonably self-explanatory. Please feel free to contact me with issues regarding the documentation; it is very informative and will help improve documentation, if I understand where users get stuck. However, I am of course NOT a help desk for experimentation or issues with using R or the R-Commander.
Ulrike Groemping
Groemping, U. (2011). Tutorial for designing experiments using the R package RcmdrPlugin.DoE.
Report 4/2011, Reports in Mathematics, Physics and Chemistry, Department II,
Beuth University of Applied Sciences Berlin.
The report itself: http://www1.bht-berlin.de/FB_II/reports/Report-2011-004.pdf.
Example data: https://prof.bht-berlin.de/fileadmin/prof/groemp/downloads/Tutorial_DoEGUI_exampleData.zip.
See also DoE.base-package,
FrF2-package,
DoE.wrapper-package for the packages behind
this R-Commander plugin.