The HPSVM Procedure

For the interior point method, the scoring process is straightforward. As long as the training weight parameters and bias are known, the scoring process is just a linear combination. The event or nonevent is decided by the decision function. The decision function variable _P_ is defined in both the training output table and the scored data set. If the value of _P_ is less than or equal to 0, then the prediction is an event; otherwise, it is a nonevent.

For the interior point method, the score code is provided for a linear kernel and for a polynomial kernel of degree 2 and 3.

For the active-set method, three data sets (Outclass, Outfit, and Outest) are generated, and the SVMSCORE procedure is used for data scoring. See Active-Set Method. For more information about the SVMSCORE procedure, see SAS Enterprise Miner and SAS Text Miner Procedures Reference for SAS.