Class LibLinearClassificationTrainer

java.lang.Object
org.tribuo.common.liblinear.LibLinearTrainer<Label>
org.tribuo.classification.liblinear.LibLinearClassificationTrainer
All Implemented Interfaces:
com.oracle.labs.mlrg.olcut.config.Configurable, com.oracle.labs.mlrg.olcut.provenance.Provenancable<org.tribuo.provenance.TrainerProvenance>, WeightedLabels, org.tribuo.Trainer<Label>
public class LibLinearClassificationTrainer extends LibLinearTrainer<Label> implements WeightedLabels
A Trainer which wraps a liblinear-java classifier trainer.

See: 

 Fan RE, Chang KW, Hsieh CJ, Wang XR, Lin CJ.
 "LIBLINEAR: A library for Large Linear Classification"
 Journal of Machine Learning Research, 2008.
and for the original algorithm: 
 Cortes C, Vapnik V.
 "Support-Vector Networks"
 Machine Learning, 1995.

  • Constructor Details

    • LibLinearClassificationTrainer

      public LibLinearClassificationTrainer()
      Creates a trainer using the default values (LinearClassificationType.LinearType.L2R_L2LOSS_SVC_DUAL, 1, 0.1, Trainer.DEFAULT_SEED).

    • LibLinearClassificationTrainer

      public LibLinearClassificationTrainer(LinearClassificationType trainerType, double cost, double terminationCriterion)
      Creates a trainer for a LibLinearClassificationModel.

      Uses Trainer.DEFAULT_SEED as the RNG seed. Sets maxIterations to 1000.

      Parameters:
      trainerType - Loss function and optimisation method combination.
      cost - Cost penalty for each incorrectly classified training point.
      terminationCriterion - How close does the optimisation function need to be before terminating that subproblem (usually set to 0.1).

    • LibLinearClassificationTrainer

      public LibLinearClassificationTrainer(LinearClassificationType trainerType, double cost, int maxIterations, double terminationCriterion)
      Creates a trainer for a LibLinear model

      Uses Trainer.DEFAULT_SEED as the RNG seed.

      Parameters:
      trainerType - Loss function and optimisation method combination.
      cost - Cost penalty for each incorrectly classified training point.
      maxIterations - The maximum number of dataset iterations.
      terminationCriterion - How close does the optimisation function need to be before terminating that subproblem (usually set to 0.1).

    • LibLinearClassificationTrainer

      public LibLinearClassificationTrainer(LinearClassificationType trainerType, double cost, int maxIterations, double terminationCriterion, long seed)
      Creates a trainer for a LibLinear model
      Parameters:
      trainerType - Loss function and optimisation method combination.
      cost - Cost penalty for each incorrectly classified training point.
      maxIterations - The maximum number of dataset iterations.
      terminationCriterion - How close does the optimisation function need to be before terminating that subproblem (usually set to 0.1).
      seed - The RNG seed.

  • Method Details