1.   Data files (4 folds)

Au6Classa.mat
Au6Classb.mat
Au6Classc.mat
Au6Classd.mat

Each MATLAB-format file contains :
   X:  900 training samples (6 classes X 150) 
   Xv: 300 validation (test) samples (6 classes X 50)
   tch : one-hot encoded labels for training data
   tchv: one-hot encoded labels for test data

   Data is log-spectrograms 40 freq X 24 time (960 dimension) 


For instructions on how to recreate published results using
PBN Toolkit, contact author:

p.m.baggenstoss@ieee.org