Software

This page gives access to PRTools and will list other toolboxes based on PRTools. They can be downloaded for free. Many of them are in fact a trial version and will have some restrictions w.r.t. dataset sizes or otherwise.

• PRTools4, Pattern Recognition Tools: about 300 user routines for preprocessing, feature extraction, transformations, density estimation, classification and evaluation. See the PRTools introduction page for a more complete description. There are also other toolboxes available that can be freely used for non-commercial research purposes. The PRTools distribution contains mainly mfiles, partially pcode. [download] [user guide]

• PRTools5. This has recently been developed. It works well with the simultaneous use of the Matlab Statistical Toolbox Stats and integrates a number of its classifiers. Transition from PRTools4 to PRTools5 is smooth. See the transition notes for more information. [download] [user guide][examples]

• PRTools for Octave. Octave is a public domain Matlab clone. It has some limitations, is slower than Matlab, but can still be very useful. There is a trial version of PRTools for Octave available. [download] [remarks]. The recent versions of PRTools5, however, also works properly under Octave. Thereby the Octave trial version will not be updated in the future.

• ClusterTools, a toolbox for cluster analysis containing traditional algorithms like KMeans, KCentres and various hierarchical clustering schemes, as well as more advanced algorithms like MeanShift, KNN-modeseeking and Exemplar. It is focused on multi-level (multi-scale) clustering and uses labeled datasets for evaluation. [download] [user guide].

• DisTools, Dissimilarity based pattern recognition: computation of dissimilarities, transformations, embedding, dissimilarity space, classification and evaluation. See the DisTools page for a more extensive description. [download] [user guide] [Introductory Examples] [Dissimilarity Representation Course]

• DD_Tools. Data Description toolbox created by David Tax. It contains many routines for one-class classification. [download] [introduction] [manual]

• MIL_Tools. Multiple Instance toolbox created by David Tax and Veronika Cheplygina. This preliminary toolbox contains algorithms to train, investigate, visualize and evaluate multiple instance learning classifiers. [download] [introduction].