A B C D E F G H I J L M N P R S T U V

C

CIPFP - Class in umbc.ebiquity.BayesOWL.coreAlgorithms

Q(X): A joint distribution of random variables in X.

CIPFP(JointProbDistribution, ProbDistribution[]) - Constructor for class umbc.ebiquity.BayesOWL.coreAlgorithms.CIPFP

Constructor.

CIPFPOneR - Class in umbc.ebiquity.BayesOWL.coreAlgorithms

This class implements algorithm "CIPFP" for one constraint.

Q(X): A joint distribution of random variables in X.
R(Si|Li): A conditional constraint to be satisfied and Si, Li are disjoint non-empty subsets of X.
(1) Q_k(X) = 0 if Q_k-1(Si|Li) = 0
(2) Q_k(X) = Q_k-1(X) * R(Si|Li) / Q_k-1(Si|Li) if Q_k-1(Si|Li) > 0
Assume Q(X), R(Si|Li) are valid distributions, complete and consistent.

CIPFPOneR(JointProbDistribution, CondProbDistribution) - Constructor for class umbc.ebiquity.BayesOWL.coreAlgorithms.CIPFPOneR

Constructor.

com.swtdesigner - package com.swtdesigner

computation() - Method in class umbc.ebiquity.BayesOWL.coreAlgorithms.CIPFPOneR

The computation process of one-step conditional iterative proportional fitting procedure (CIPFP), for a single conditional constraint.

computation() - Method in class umbc.ebiquity.BayesOWL.coreAlgorithms.DIPFPConditionalOneR

The computation process of one-step de-composed iterative proportional fitting procedure (DIPFP), for a single conditional constraint, either local or non-local.

computation() - Method in class umbc.ebiquity.BayesOWL.coreAlgorithms.DIPFPMarginalOneR

The computation process of one-step de-composed iterative proportional fitting procedure (DIPFP), for a single marginal constraint, either local or non-local.

computation() - Method in class umbc.ebiquity.BayesOWL.coreAlgorithms.IPFPOneR

The computation process of one-step iterative proportional fitting procedure (IPFP) for a single constraint.

computation() - Method in class umbc.ebiquity.BayesOWL.coreAlgorithms.SDIPFPConditionalOneR

The computation process of one-step simplified de-composed iterative proportional fitting procedure (SDIPFP), for a single simple local conditional constraint.

computation() - Method in class umbc.ebiquity.BayesOWL.coreAlgorithms.SDIPFPMarginalOneR

The computation process of one-step simplified de-composed iterative proportional fitting procedure (SDIPFP), for a single simple local marginal constraint.

computation() - Method in class umbc.ebiquity.BayesOWL.coreAlgorithms.SDIPFPOneR

The computation process of one-step simplified de-composed iterative proportional fitting procedure (SDIPFP), for a single simple constraint.

ConditionalConstraint - Class in umbc.ebiquity.BayesOWL.commonDefine

ConditionalConstraint is an abstract class.

ConditionalConstraint(String, CondProbDistribution) - Constructor for class umbc.ebiquity.BayesOWL.commonDefine.ConditionalConstraint

Constructs a conditional constraint with the specific type, i.e., local or non-local.

CondProbDistribution - Class in umbc.ebiquity.BayesOWL.commonDefine

This class implements a full conditional probability distribution, which includes:
(1) An array of n prior random variables {vi}, i = 1 to n, each vi has vdi states
(2) An array of m condition random variables {ci}, i = 1 to m, each ci has cdi states
(3) A m+n-dimensional array with size "cd1 x cd2 x ...

CondProbDistribution(RandomVariable[], RandomVariable[]) - Constructor for class umbc.ebiquity.BayesOWL.commonDefine.CondProbDistribution

Constructs a new conditional probability distribution table, given a set of prior and condition variables.

CondProbDistribution(CondProbDistribution) - Constructor for class umbc.ebiquity.BayesOWL.commonDefine.CondProbDistribution

Copy Constructor.

Constraint - Class in umbc.ebiquity.BayesOWL.commonDefine

Constraint is an abstract class.

Constraint(String) - Constructor for class umbc.ebiquity.BayesOWL.commonDefine.Constraint

Constructor.

constructBN(String[], ExNode.TAG[], String[][]) - Method in class umbc.ebiquity.BayesOWL.constructor.BNConstructor

Method to construct BN structure.

containsCondVariable(String) - Method in class umbc.ebiquity.BayesOWL.commonDefine.CondProbDistribution

This method tests whether the given random variable is involved in the "condition" part of this conditional probability distribution.

containsPriorVariable(String) - Method in class umbc.ebiquity.BayesOWL.commonDefine.CondProbDistribution

This method tests whether the given random variable is involved in the "prior" part of this conditional probability distribution.

containsVariable(String) - Method in class umbc.ebiquity.BayesOWL.commonDefine.JointProbDistribution

This method tests whether the given random variable is involved in this joint probability distribution.

CPTConstructor - Class in umbc.ebiquity.BayesOWL.constructor

CPTConstructor implements BayesOWL's Conditional Probability Table Constructor.
It takes probability constraints and a target BN as inputs.
To have input constraints, go to see class umbc.ebiquity.BayesOWL.commonDefine.Constraint.java

CPTConstructor(Net, Constraint[]) - Constructor for class umbc.ebiquity.BayesOWL.constructor.CPTConstructor

Constructor.

CrossEntropy - Class in umbc.ebiquity.BayesOWL.commonMethod

This class provides method to compute the cross entropy between two joint probability distributions.

Definition of I-divergence (also named Kullback-Leibler divergence, cross-entroy):
if P(X)< I(P||Q) = sum_over_all_X's assignments{P(x)log(P(x)/Q(x))}, only picks those assignments when P(x)>0
if P(X)!(<<)Q(X), then:
I(P||Q) = positive infinity

Generally,I(P||Q) != I(Q||P)
Note: When P is dominated by Q, we have: 0/0 = 0 and 0*log(0/0) = 0.
see also:
http://en.wikipedia.org/wiki/Kullback-Leibler_divergence
http://en.wikipedia.org/wiki/Cross_entropy
Assume: P(X) and Q(X) are valid distributions, w.r.t our implementation here.

CrossEntropy(JointProbDistribution, JointProbDistribution) - Constructor for class umbc.ebiquity.BayesOWL.commonMethod.CrossEntropy

Constructs a new class for cross entropy.