The addmtoolbox package provides a set of functions that facilitate the process of fitting (attentional) drift diffusion models to experimental data.
The basic drift diffusion process is implemented in a highly optimized c++ function, which is complemented with a host of R functions that help with data preparation, facilitation of parameter optimization and parallelization.
The package is in an early development stage, but basic functionality is in working condition. Make sure to read the Documentation section for info how to use the package. Content that is mentioned in the package vignettes is currently the safest to run.
Having that said, the package will improve dramatically in the next weeks/month so don't be put off by a few bugs in the meanwhile.
The following guide is specifically for Mac OS systems. Windows systems have not yet been tested. On Linux/Ubuntu system, directly follow up with step: "Installing addmtoolbox from github", as gsl is often/always included in the distribution as well as all utilities that you gain on Macs by installing Xcode. This guide will carry you through the complete installation process, starting from the installation of the R progamming language on your system.
Installation relevant dependencies
The addmtoolbox package depends notably on the RcppZiggurat package which in turn depends on the RcppGSL package. For the RcppGSL package to work, we need to install the GNU GSL scientific library.
First, you need to have Xcode installed on your system, which if not present can be downloaded for free from the Apple Appstore. After installing Xcode, you also need the command line tools from Apple (Apple wants you to sign in to download those).
Installing GSL Library
Below is a step by step guide to installing the GNU GSL scientific library.
- First download the current version and unpack it if not done automatically. (choose the file which has the highest version number: gsl-1.16.tar.gz at the time of writing)
- Go to the terminal and navigate to the directory to which you downloaded the gsl library. (useful terminal commands: cd, ls)
./configurewait until the the execution finished and then type
maketo build the library. (you can find these instructions in the INSTALL file inside of the downloaded gsl folder). The build process takes a while now.
- Lastly type
make install, which will finally install the library on your system. It is likely that you have to use
sudo make install, as the default folder which the library is installed in requires superuser access.
Please proceed only if you have successfully installed Xcode and the command line tools. If you install R without these things present on you system, there is some risk that the addmtoolbox will not run/install.
Given that you followed the instructions so far, we are now ready to get the R programming language ready for usage on our system (if not present). The package is built under R Version 3.1.2. and I strongly suggest to install this or (if present) a later version.
We can proceed by opening RStudio.
Installing addmtoolbox from github
To install the addmtoolbox package from github, first the devtools package must be loaded in R.
This can be easily done by typing
install.packages("devtools") in the R Console to install the package and the typing
library(devtools) to load it.
install_github('AlexanderFengler/addmtoolbox', build_vignettes=TRUE) to install the addmtoolbox package and make it available in your R Session with
You should be able to use the provided functions now.
Besides the standard documentation you will find in the help window in RStudio, the package come with vignettes, one of which carries you through a complete model-fit.
Be sure to include the
build_vignettes = TRUE parameter when installing the package, otherwise the following instruction will not work.
To list the vignettes of the addmtoolbox package, type
vignette(package = "addmtoolbox"). This lets you see the names of the vignette-documents that come with package. Most importantly you are going to see addmtoolbox_modelfit_walkthrough vignette.
To open the vignette, type
vignette("addmtoolbox_general_info"). This vignette will carry you through a model-fit to get you started with the package.
Below a list of R packages that are used in the source code. In case you would like to really understand and contribute to the codebase, I think it will make it considerably easier if you first familiarize yourself with the basic utilities these packages offer (in case you have not already), as they are used quite heavily.