Module es.upm.fi.cig.multictbnc
Class BNLogLikelihood
java.lang.Object
es.upm.fi.cig.multictbnc.learning.structure.optimisation.scores.AbstractLikelihood
es.upm.fi.cig.multictbnc.learning.structure.optimisation.scores.bn.BNLogLikelihood
- All Implemented Interfaces:
BNScoreFunction
Implements the log-likelihood score for Bayesian networks with nodes that have CPTs.
-
Constructor Summary
ConstructorDescriptionBNLogLikelihood
(String penalisationFunction) Receives the name of the penalisation function for the structure complexity. -
Method Summary
Modifier and TypeMethodDescriptiondouble
Computes the (penalised) log-likelihood score for a discrete Bayesian network.Gets an identifier for the score function.Methods inherited from class es.upm.fi.cig.multictbnc.learning.structure.optimisation.scores.AbstractLikelihood
getNamePenalisationFunction, getPenalisationFunction
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
Methods inherited from interface es.upm.fi.cig.multictbnc.learning.structure.optimisation.scores.bn.BNScoreFunction
getNamePenalisationFunction
-
Constructor Details
-
BNLogLikelihood
Receives the name of the penalisation function for the structure complexity.- Parameters:
penalisationFunction
- name of the penalisation function
-
-
Method Details
-
compute
Computes the (penalised) log-likelihood score for a discrete Bayesian network. This is done by computing the marginal log-likelihood of the graph. A uniform prior structure is assumed.- Specified by:
compute
in interfaceBNScoreFunction
- Parameters:
bn
- a Bayesian network- Returns:
- penalised log-likelihood score
-
getIdentifier
Description copied from interface:BNScoreFunction
Gets an identifier for the score function.- Specified by:
getIdentifier
in interfaceBNScoreFunction
- Returns:
- identifier for the score function.
-