SDREJECT   Add reject option to a trained pipeline

     PR=SDREJECT(P,DATA,options)

 INPUT
    P       Trained pipeline
    DATA    Training data set used to adjust rejection threshold

 OUTPUT
    PR      Pipeline P with added decision operating point

 OPTIONS
    'reject',N   Reject fraction or number of samples (optional, default=0)
    'non-target',name  Decision string of rejected samples (default: 'reject')

 DESCRIPTION
 SDREJECT adds rejection to a trained pipeline P based on data set DATA
 (typically the same as used for training the pipeline).
 By default, SDREJECT adjusts the reject threshold so that no samples
 are rejected. Using 'reject' option, we may specify the fraction or number of
 samples to be rejected.

 READ MORE
http://perclass.com/doc/guide/ROC_analysis.html#reject

sdreject is referenced in examples:

• kb27: Upgrading to perClass 4
• kb18: How to protect a trained discriminant against outliers?