Class MLPlanSKLearnBuilder

  • All Implemented Interfaces:
    ai.libs.mlplan.core.IMLPlanBuilder<ai.libs.jaicore.ml.scikitwrapper.ScikitLearnWrapper,​MLPlanSKLearnBuilder>, org.api4.java.common.control.ILoggingCustomizable

    public class MLPlanSKLearnBuilder
    extends ai.libs.mlplan.core.AbstractMLPlanBuilder<ai.libs.jaicore.ml.scikitwrapper.ScikitLearnWrapper,​MLPlanSKLearnBuilder>
    • Field Summary

      • Fields inherited from class ai.libs.mlplan.core.AbstractMLPlanBuilder

        DEFAULT_PERFORMANCE_MEASURE, DEFAULT_SEARCH_NUM_MC_ITERATIONS, DEFAULT_SEARCH_TRAIN_FOLD_SIZE, DEFAULT_SELECTION_NUM_MC_ITERATIONS, DEFAULT_SELECTION_TRAIN_FOLD_SIZE, factoryForPipelineEvaluationInSearchPhase, factoryForPipelineEvaluationInSelectionPhase
    • Constructor Summary

      Constructors 
      Constructor Description
      MLPlanSKLearnBuilder()
      Creates a new ML-Plan Builder for scikit-learn.
      MLPlanSKLearnBuilder​(boolean skipSetupCheck)
      Creates a new ML-Plan Builder for scikit-learn.
    • Method Summary

      All Methods Instance Methods Concrete Methods 
      Modifier and Type Method Description
      MLPlanSKLearnBuilder getSelf()  
      MLPlanSKLearnBuilder withPreferredComponentsFile​(java.io.File preferredComponentsFile, java.lang.String preferableCompnentMethodPrefix)
      Creates a preferred node evaluator that can be used to prefer components over other components.
      MLPlanSKLearnBuilder withUnlimitedLengthPipelineSearchSpace()
      Configures ML-Plan to use the search space with unlimited length preprocessing pipelines.
      • Methods inherited from class ai.libs.mlplan.core.AbstractMLPlanBuilder

        build, build, checkPreconditionsForInitialization, getAlgorithmConfig, getCandidateEvaluationTimeOut, getComponentParameterConfigurations, getComponents, getDataset, getHASCOFactory, getLearnerEvaluationFactoryForSearchPhase, getLearnerEvaluationFactoryForSelectionPhase, getLearnerFactory, getLoggerName, getNodeEvaluationTimeOut, getPortionOfDataReservedForSelectionPhase, getRequestedInterface, getSearchEvaluatorFactory, getSearchSelectionDatasetSplitter, getSearchSpaceConfigFile, getSelectionEvaluatorFactory, getTimeOut, prepareNodeEvaluatorInFactoryWithData, setLoggerName, withAlgorithmConfig, withAlgorithmConfigFile, withCandidateEvaluationTimeOut, withClassifierFactory, withDataset, withDatasetSplitterForSearchSelectionSplit, withMCCVBasedCandidateEvaluationInSearchPhase, withMCCVBasedCandidateEvaluationInSelectionPhase, withNodeEvaluationTimeOut, withNumCpus, withPipelineValidityCheckingNodeEvaluator, withPortionOfDataReservedForSelection, withPreferredNodeEvaluator, withRandomCompletionBasedBestFirstSearch, withRequestedInterface, withSearchFactory, withSearchPhaseEvaluatorFactory, withSearchSpaceConfigFile, withSeed, withSelectionPhaseEvaluatorFactory, withTimeOut
      • Methods inherited from class java.lang.Object

        clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
    • Constructor Detail

      • MLPlanSKLearnBuilder

        public MLPlanSKLearnBuilder()
                             throws java.io.IOException
        Creates a new ML-Plan Builder for scikit-learn.
        Throws:
        java.io.IOException - Thrown if configuration files cannot be read.
      • MLPlanSKLearnBuilder

        public MLPlanSKLearnBuilder​(boolean skipSetupCheck)
                             throws java.io.IOException
        Creates a new ML-Plan Builder for scikit-learn.
        Parameters:
        skipSetupCheck - Flag whether to skip the system's setup check, which examines whether the operating system has python installed in the required version and all the required python modules are installed.
        Throws:
        java.io.IOException - Thrown if configuration files cannot be read.
    • Method Detail

      • withUnlimitedLengthPipelineSearchSpace

        public MLPlanSKLearnBuilder withUnlimitedLengthPipelineSearchSpace()
                                                                    throws java.io.IOException
        Configures ML-Plan to use the search space with unlimited length preprocessing pipelines.
        Returns:
        The builder object.
        Throws:
        java.io.IOException - Thrown if the search space configuration file cannot be read.
      • withPreferredComponentsFile

        public MLPlanSKLearnBuilder withPreferredComponentsFile​(java.io.File preferredComponentsFile,
                                                                java.lang.String preferableCompnentMethodPrefix)
                                                         throws java.io.IOException
        Creates a preferred node evaluator that can be used to prefer components over other components.
        Parameters:
        preferredComponentsFile - The file containing a priority list of component names.
        preferableCompnentMethodPrefix - The prefix of a method's name for refining a complex task to preferable components.
        Returns:
        The builder object.
        Throws:
        java.io.IOException - Thrown if a problem occurs while trying to read the file containing the priority list.