prtDecision Base class for sll prt decision objects
 
   This is a base class from which all prt decision objects inherit. It
   cannot be instantiated. prtDecision objects inherit the RUN function
   from prtAction objects. 
 
  prtDecision objects accept the outputs of classifiers or clusterers
  and return class or cluster labels.  The output of a prtDecision 
  object's run function is always a nObservations x 1 data set where
  the observations are integer valued and specify the hypothesized
  class for each observation
 
  Decisions can be used in two ways - either as part of an algorithm:
 
  algo = prtClassKnn + prtDecisionBinaryMinPe;
 
  Or as the internalDecider object inside a prtClass object:
 
   myKnn = prtClassKnn;
   myKnn.internalDecider = prtDecisionBinaryMinPe;
   myKnn = myKnn.train(dsTrain);
  
  In the first case, the resulting object is a prtAlgorithm, which
  changes the behavior of PLOT.  In the second case, the decision is
  incorporated into the classifier, allowing the decision contours to
  be plotted easily.
 
  Note that the regardless of how one uses a decider, the outputs of
  the RUN function should be identical.
 
   % Example 1:
  
  % This example demonstrates use of a prtDecision object to perform
  % class labeling as part of a prtAlgorithm
 
  dsTrain = prtDataGenUnimodal;
  dsTest = prtDataGenUnimodal;
  algo = prtClassKnn + prtDecisionBinaryMinPe;
  algo = algo.train(dsTrain);
  plot(algo); title('Algorithm implementation');
  outAlgo = algo.run(dsTest);
 
   % Example 2:
  
   % This example demonstrtes the use of a prtDecision objet to perform
   % class labeling as prtClass objects internalDecider
 
   myKnn = prtClassKnn;
   myKnn.internalDecider = prtDecisionBinaryMinPe;
   myKnn = myKnn.train(dsTrain);
   outIntDec = myKnn.run(dsTest);
   figure; plot(myKnn); title('internalDecider implementation');