The routine DSCRM allows linear or quadratic discrimination and the use of either reclassification, split sample, or the leaving‑out‑one methods in order to evaluate the rule. Moreover, DSCRM can be executed in an online mode, that is, one or more observations can be added to the rule during each invocation of DSCRM.

The mean vectors for each group of observations and an estimate of the common covariance matrix for all groups are input to DMSCR. These estimates can be computed via routine DSCRM. Output from DMSCR are linear combinations of the observations, which at most separate the groups. These linear combinations may subsequently be used for discriminating between the groups. Their use in graphically displaying differences between the groups is possibly more important, however.

Nearest neighbor discrimination is performed in routine NNBRD. In this routine, the user can set the number of nearest neighbors to be used in the discrimination and the threshold for classification. Split samples can also be used.