mdatools is a free open source toolbox for analysis of multivariate experimental data based on methods widespread in Chemometrics. The toolbox can work with conventional two-way datasets (where rows are objects or measurements and columns are variables) as well as with colour and hyperspectral images. It has both command line and GUI tools allowing to explore models and datasets interactively. The project is currently under developing and e.g. GUI is in alpha version and has limited functionality, however most of the command line methods work well in the current version.
It is recommended to start learning the toolbox with a guide, which gives
a brief overview of basic methods with a lot of examples (web mdatools_quick.html).
It allows you to start working with the toolbox after 30 minutes
learning. Then just continue with User guide (web mdatools.html) to get acquainted with all
available features and details.
The toolbox was tested using MATLAB 2012b and 2014b and should also work on versions in between. For the basic functionality you do not need any additional toolboxes installed, however for GUI tools Image Processing toolbox is needed (this will be fixed in future releases). All methods were tested using well-known datasets, but there still could be some bugs, in this case please report to [email protected].
The current version is 0.0.13 (from 21.05.2015).
Installation is easy and the procedure is similar to most of the others 3rd party MATLAB Toolboxes.
Download zip archive for the latest release and unzip it to the folder you use to keep you MATLAB code,
for example to Documents/MATLAB or Documents/MATLAB/Toolboxes. Open the folder in MATLAB
environment and run install.m script, which will do the rest.
If you upgrade from previous version, run uninstall on the old version first (it will clear the path
list and remove folder with toolbox), then unzip the new version and follow the instructions above.
If you find a bug, please, send a message to [email protected] with detailed description or use a form on GitHub. Any comments and suggestions will also be highly appreciated.