Class BayesNetworkScore
java.lang.Object
de.uka.ipd.sdq.dsexplore.bayesnets.utility.BayesNetworkScore
Class for calculating the score of a Bayesian Network. It
contains various scoring techniques which the user can use at
his/her own discretion.
-
Field Summary
-
Constructor Summary
ConstructorDescriptionBayesNetworkScore
(int[][] GraphMatrix, int[][] DataMatrix) Constructor for the class. -
Method Summary
Modifier and TypeMethodDescriptiondouble
AIC()
Scores the network according to the Akaike Information Criterion (AIC) technique.double
Scores the network according to the BDeu scoring technique.double
BIC()
Scores the network according to the Bayesian Information Criterion (BIC) technique.double
Scores the network according to the K2 scoring techniquedouble
LogLik()
Scores the network using the log-likelihood technique.static void
-
Field Details
-
SQTPI
protected static final double SQTPI- See Also:
-
-
Constructor Details
-
BayesNetworkScore
public BayesNetworkScore(int[][] GraphMatrix, int[][] DataMatrix) Constructor for the class. Initializes the Graph and Data fields- Parameters:
int
- [][] GraphMatrix - Adjacency matrix for the Graphint
- [][] DataMatrix - Matrix for the Data containing 0 and 1.
-
-
Method Details
-
main
-
K2NetworkScore
public double K2NetworkScore()Scores the network according to the K2 scoring technique- Parameters:
No
- parameters- Returns:
- The score of the network as a double
-
BDeuNetworkScore
public double BDeuNetworkScore()Scores the network according to the BDeu scoring technique. (Not recommended for use currently)- Parameters:
No
- parameters- Returns:
- The score of the network as a double
-
LogLik
public double LogLik()Scores the network using the log-likelihood technique.- Parameters:
No
- parameters- Returns:
- The score of the network as a double
-
BIC
public double BIC()Scores the network according to the Bayesian Information Criterion (BIC) technique.- Parameters:
No
- parameters- Returns:
- The score of the network as a double
-
AIC
public double AIC()Scores the network according to the Akaike Information Criterion (AIC) technique.- Parameters:
No
- parameters- Returns:
- The score of the network as a double
-