coreComponents
Interface NonExchangeableDistance

All Known Implementing Classes:
ValueDifferenceMetric

public interface NonExchangeableDistance

This interface is useful only for classifiers which require the exchangeability assumption to be made about the data - i.e. Confidence and Probability classifiers.

Version:
$Revision: 1.1$
Author:
David Lindsay (davidl@cs.rhul.ac.uk)

Method Summary
 double calculateTentativeDistance(Instance inst1, Instance inst2)
          Calculate the tentative instance of the classifier.
 boolean setTentativeInstance(Instance inst)
          Sets the instances which we tentatively update the distance metric for.
 

Method Detail

setTentativeInstance

public boolean setTentativeInstance(Instance inst)
Sets the instances which we tentatively update the distance metric for.


calculateTentativeDistance

public double calculateTentativeDistance(Instance inst1,
                                         Instance inst2)
Calculate the tentative instance of the classifier.



Copyright (c) 2003 David Lindsay, Computer Learning Research Centre, Dept. Computer Science, Royal Holloway, University of London