Preamble

• ECOC PAK C++ supports MEXfiles MATLAB interfaces for the most of the ECOC PAK C++ library's functionalities.


• The source code for the MEXfiles is available inside the mex/ folder inside the current version's tarbal.

---

Reference

ecocpak() · ecocpakcv() · mat2libsvm() · libsvm2mat()


• ecocpak()
  
  
  • Usage:
    
    [error_rate] = ecocpak(training_samples, training_labels, testing_samples, testing_labels,...)
  
  
  • Description:
    
    Training - Testing procedure for ecocpak.
  
  
  • For a full description run ecocpak with no arguments.
  
  
• ecocpakcv()
  
  
  • Usage:
    
    [error_rate] = ecocpakcv(training_samples, training_labels,...)
  
  
  • Description:
    
    K-fold cross-validation procedure for ecocpak.
  
  
  • For a full description run ecocpakcv with no arguments.
  
  
• mat2libsvm()
  
  
  • Usage:
    
    mat2libsvm(samples, labels, output_file_name)
  
  
  • Description:
    
    The mat2libsvm() function transforms a dataset from MATLAB (i.e., a matrix with the sample feature vectors and a vector with their respective class labels) to LIBSVM format file.
  
  
  • Mandatory Input Arguments:
    
    
    • samples : A 2D matrix where its rows denote the sample feature vectors of the dataset
    • labels : An 1D vector with the class labels
    • output_file_name : A string with the desired output file name
  
  
  • Example:
    
    mat2libsvm(iris_samples, iris_labels, 'iris_dataset')


• libsvm2mat()
  
  
  • Usage:
    
    [samples, labels] = libsvm2mat(input_file_name)
  
  
  • Description:
    
    The libsvm2mat() function transforms a dataset from LIBSVM format file to a MATLAB 2D matrix where its rows represent the dataset's feature vectors and an 1D vector with the dataset's feature vectors respective class labels.
  
  
  • Mandatory Input Arguments:
    
    
    • input_file_name : A string with the desired input file name
  
  
  • Example:
    
    [iris_samples, iris_labels] = libsvm2mat('iris_dataset')