DMLT is a Machine Learning toolbox written in Matlab and C. This toolbox is developed at the Donders Institute for Brain, Cognition and Behaviour and provides a general interface to support the integration of new statistical machine learning methods by writing high level wrappers. It allows complex methods to be built from simple building blocks and makes the use of cross-validation and permutation testing as easy as writing one line of Matlab code. The code requires at least Matlab distribution 184.108.40.2064 (R2008a).
Most functions in this toolbox are licensed under the GNU General Public License (GPL), see http://www.gnu.org for details. Unauthorised copying and distribution of functions that are not explicitely covered by the GPL is not allowed.
The most recent version of DMLT can be downloaded as a zipfile at https://github.com/marcelge/DMLT/zipball/master
It is recommended though to use install DMLT by cloning the repository:
git clone git://github.com/distrep/DMLT.git
or, if you have write privileges, using:
git clone email@example.com:distrep/DMLT.git
The repository may then always be updated to the latest version by issuing the
git pull command in the toolbox root folder.
DMLT documentation is added automatically to the Matlab help facility when adding DMLT to the search path, e.g. using:
assuming you start your Matlab session in the DMLT root folder. You may also browse html/guide.html for a quick guide on how to use DMLT.