• Matlab Pattern Recognition Toolbox for representation and generalization


  • More than 300 dedicated pattern recognition routines
  • About 500 support routines
  • representation, feature extraction and selection
  • pre-processing of raw data files included in the toolbox,
  • integrated handling of data in files and data in RAM
  • linear and non-linear transformations
  • crisp, soft and target labels
  • multi-labeling system
  • more than 30 different classifiers
  • large set of combining classifiers
  • crossvalidation, learning curves, feature curves, EOC
  • integration of various Mathworks routines for statistics and machine learning
  • runs (with some restrictions) under Octave

Typical Usage

  • prototyping of recognition systems including preprocessing
  • the analysis of new data by standard tools
  • comparative studies for evaluating new, advanced tools
  • student courses