Module es.upm.fi.cig.multictbnc
Package es.upm.fi.cig.multictbnc.learning.structure

Class StructureLearningAlgorithmFactory

java.lang.Object
es.upm.fi.cig.multictbnc.learning.structure.StructureLearningAlgorithmFactory
public class StructureLearningAlgorithmFactory extends Object
Builds the specified structure learning algorithms for Bayesian networks and continuous-time Bayesian networks.

  • Constructor Details

    • StructureLearningAlgorithmFactory

      public StructureLearningAlgorithmFactory()

  • Method Details

    • getAlgorithmBN

      public static StructureLearningAlgorithm getAlgorithmBN(String algorithm, Map<String,String> param)
      Builds the specified structure learning algorithm for Bayesian networks. Current parameters: (1) scoreFunction: name of the score function (score-based/hybrid algorithms) (2) penalisationFunction: name of the penalisation function (score-based/hybrid algorithms) (3) numRestarts: number of restarts for the random-restart hill climbing (random-restart hill climbing) (3) tabuListSize: size of the tabu list (tabu search) (4) significancePC: significance of the PC algorithm (constraint-based/hybrid algorithms)
      Parameters:
      algorithm - name of the structure learning algorithm
      param - map containing the necessary parameters for the requested algorithm
      Returns:
      structure learning algorithm for Bayesian networks

    • getAlgorithmCTBN

      public static StructureLearningAlgorithm getAlgorithmCTBN(String algorithm, Map<String,String> param)
      Builds the specified structure learning algorithm for continuous-time Bayesian networks.

      (1) scoreFunction: name of the score function (score-based/hybrid algorithms) (2) penalisationFunction: name of the penalisation function (score-based/hybrid algorithms) (3) numRestarts: number of restarts for the random-restart hill climbing (Random-restart hill climbing) (3) tabuListSize: size of the tabu list (Tabu search) (4) sigTimeTransitionHyp: significance of the PC algorithm (constraint-based/hybrid algorithms) (5) sigStateToStateTransitionHyp: significance of the PC algorithm (constraint-based/hybrid algorithms) (6) maxSizeSepSet: (Hybrid algorithm)

      Parameters:
      algorithm - name of the structure learning algorithm
      param - map containing the necessary parameters for the requested algorithm
      Returns:
      structure learning algorithm for continuous-time Bayesian networks

    • getAvailableLearningMethods

      public static List<String> getAvailableLearningMethods()
      Returns the name of available optimisation methods.
      Returns:
      optimisation methods