it.unibo.cs.belief
Class Belief

java.lang.Object
  extended by it.unibo.cs.belief.Belief

public class Belief
extends java.lang.Object

This class provides some implementations of Dempster's combination rule for mass potentials (currently only one implementation is available).

Version:
0.1
Author:
Matteo Magnani

Constructor Summary
private Belief()
           
 
Method Summary
static MassPotential combine(MassPotential[] pot)
          A direct implementation of Dempster's rule of combination for many mass potentials.
static MassPotential combine(MassPotential m1, MassPotential m2)
          A direct implementation of Dempster's rule of combination for two mass potentials m1 and m2.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

Belief

private Belief()
Method Detail

combine

public static MassPotential combine(MassPotential m1,
                                    MassPotential m2)
A direct implementation of Dempster's rule of combination for two mass potentials m1 and m2. This method combines two mass potentials by directly multiplying all pairs of propositions in F1 x F2, where F1 and F2 are respectively the focal sets of m1 and m2.

This method returns an unnormalized mass potential. Therefore, to obtain a mass function it is necessary to normalize the result. This method can combine any two mass potentials. If the corresponding mass functions have disjoint cores, i.e., their combination does not exist, an error is raised if we try to normalize the result.

Parameters:
m1 - a mass potential
m2 - a mass potential
Returns:
the unnormalized m1 (x) m2 mass potential
See Also:
MassPotential.normalize()

combine

public static MassPotential combine(MassPotential[] pot)
A direct implementation of Dempster's rule of combination for many mass potentials. This method combines N mass potentials by directly multiplying all pairs of propositions F1 x F2 x .. x FN, where F1, F2,..., FN are respectively the focal sets of m1, m2,..., mN.

This method returns an unnormalized mass potential. Therefore, to obtain a mass function it is necessary to normalize the result. This method can combine any array of mass potentials. If the corresponding mass functions have disjoint cores, i.e., their combination does not exist, an error is raised if we try to normalize the result.

Parameters:
pot - an array of mass potentials
Returns:
the unnormalized m1 (x) m2 (x) ... (x) mN mass potential
See Also:
MassPotential.normalize()