Documentation of 'edu.uci.jforests.config.TrainingConfig' Java class
TrainingConfig
edu.uci.jforests.config

Class TrainingConfig

    • Field Summary

      Fields 
      Modifier and Type Field and Description
      java.lang.String evaluationMetric
      The name of the evaluation metric to be used during training.
      java.lang.String featureNamesFilename
      If feature names are needed and they are not stored in the training file, they can be loaded from this external file.
      java.lang.String learningAlgorithm
      The name of the algorithm to be used for training.
      int numThreads
      Number of threads to use.
      boolean printIntermediateValidMeasurements
      If this parameter is set to true, the performance on validation data is printed during the iterations of training.
      int randomSeed
      The random seed to be used for training.
      static java.lang.String TRAIN_FILENAME 
      java.lang.String trainFilename
      Training file name
      double trainFraction
      If this parameter is set to a value less than 1.0, only a fraction of the training data will be used for training.
      static java.lang.String VALID_FILENAME 
      java.lang.String validFilename
      Validation file name
      double validFraction
      If this parameter is set to a value less than 1.0, only a fraction of the validation data will be used for validation.
      boolean validOutOfTrain
      If for training only a fraction of data is used and this parameter is set to true, then for validation we will use the data in training input file which is left out of training.
    • Constructor Summary

      Constructors 
      Constructor and Description
      TrainingConfig() 
    • Field Detail

      • trainFilename

        public java.lang.String trainFilename
        Training file name
      • validFilename

        public java.lang.String validFilename
        Validation file name
      • featureNamesFilename

        public java.lang.String featureNamesFilename
        If feature names are needed and they are not stored in the training file, they can be loaded from this external file. Each line in the file contains one feature name.
      • learningAlgorithm

        public java.lang.String learningAlgorithm
        The name of the algorithm to be used for training. For example, the value of "Bagging-RegressionTree" means that we want Bagging wrapped around Regression trees.
      • evaluationMetric

        public java.lang.String evaluationMetric
        The name of the evaluation metric to be used during training. Default is AUC. Other currently implemented metrics include RMSE and BalancedYoundenIndex.
      • trainFraction

        public double trainFraction
        If this parameter is set to a value less than 1.0, only a fraction of the training data will be used for training.
      • validFraction

        public double validFraction
        If this parameter is set to a value less than 1.0, only a fraction of the validation data will be used for validation.
      • validOutOfTrain

        public boolean validOutOfTrain
        If for training only a fraction of data is used and this parameter is set to true, then for validation we will use the data in training input file which is left out of training.
      • numThreads

        public int numThreads
        Number of threads to use. By default this is set to the number of processors on the machine. However, for debugging, sometimes it is needed to set it to 1.
      • randomSeed

        public int randomSeed
        The random seed to be used for training.
      • printIntermediateValidMeasurements

        public boolean printIntermediateValidMeasurements
        If this parameter is set to true, the performance on validation data is printed during the iterations of training.
    • Constructor Detail

      • TrainingConfig

        public TrainingConfig()

DMelt 3.0 © DataMelt by jWork.ORG

You see the box below because you did not login.