perClass Documentation
version 5.1 (31-May-2017)
 SDPREP - data preprocessing and feature computation

   P=SDPREP(DATA,METHOD,OPTIONS)

 INPUTS
   DATA - data set
   METHOD - method name

 OUTPUTS
   P - preprocessing pipeline

 METHODS
  'submean' - subtract sample mean
  'divsum' - divide by sum (for each sample)
  'divmean' - divide by mean (for each sample)
  'divband',B - divide by value of specific feature (band B) for each sample
  'smooth' - smooth 1D spectra (default sigma=1.0)
    'sigma',S - smoothing parameter (neighborhood defined by +/- 3*sigma)
  'der' - apply Gaussian derivative filter (def: sigma=1)
  'kernel',K - convolve 1D spectra with user-defined kernel

 Computations based on specific features (bands). A,B are feature indices.
  'a-b',A,B
  'a/b',A,B
  '(a-b)/(a+b)',A,B
  '(a+b)/(a-b)',A,B
  'a/(b*c)',A,B,C
  'a/(b-c)',A,B,C
    'add' - copy all input features and add the computed feature to the end

 EXAMPLES
  >> p=sdprep(data,'der','sigma',2); data2=data*p
 Add spectral index:
  >> p=sdprep(data,'(a-b)/(a+b)',53,107,'add'); data2=data*p