package activeml

  1. Public
  2. All

Type Members

  1. abstract class ActiveLearningAlgorithm[M, A, C <: PointLabelingContext, G <: GroupLabelingContext] extends Serializable

    An algorithm that uses active learning to iteratively train models on new labeled data.

  2. case class ActiveLearningParameters(budget: Int = 10, batchSize: Int = 5, bootstrapSize: Int = 10) extends Product with Serializable

    Parameters to run the active learning framework

  3. class ActiveLearningTrainingFuture[M] extends Awaitable[Unit]

    Future-like class for asynchronously passing back models and labeled data output by the active learning framework.

  4. class ActiveLearningTrainingState[M] extends AnyRef

  5. abstract class ActivePointSelector[M, C <: PointLabelingContext] extends Serializable

    Class for using active selection criteria to pick points for labeling during active learning.

  6. class ActiveSVMWithSGD[C <: PointLabelingContext, G <: GroupLabelingContext] extends ActiveLearningAlgorithm[SVMModel, SVMParameters, C, G]

    Active learning to train an SVM with SGD.

  7. class ForestUncertaintyFilter[C <: PointLabelingContext] extends ActivePointSelector[RandomForestModel, C]

    Uses uncertainty sampling to select points closest to the SVM Margin.

  8. class HardCodedLabelGetter extends LabelGetter[HardcodedLabelGetterPointContext, HardcodedLabelGetterGroupContext, HardcodedLabelGetterParameters]

    Dummy label getter that is initialized with labeled data and simply returns the labels when queried.

  9. case class HardcodedLabelGetterGroupContext() extends Product with Serializable

    Unused shared context for labeling groups of points.

  10. case class HardcodedLabelGetterParameters(hardcodedLabeledData: RDD[(String, LabeledPoint)]) extends Product with Serializable

    Parameters for the hardcoded label getter.

  11. case class HardcodedLabelGetterPointContext() extends Product with Serializable

    Unused context for labeling points.

  12. abstract class LabelGetter[C, G, P] extends Serializable

    A generic label-getter for adding labels to unlabeled feature vectors.

  13. case class RandomForestParameters(numTrees: Int = 100, numClasses: Int = 2, categoricalFeatureInfo: Map[Int, Int] = ..., impurity: String = "gini", maxDepth: Int = 4, maxBins: Int = 100, featureSubsetStrategy: String = "auto", algo: Algo = ...) extends Product with Serializable

  14. class SVMMarginDistanceFilter[C <: PointLabelingContext] extends ActivePointSelector[SVMModel, C]

    Uses uncertainty sampling to select points closest to the SVM Margin.

  15. case class SVMParameters(numIterations: Int = 100, stepSize: Double = 1.0, regParam: Double = 1.0, miniBatchFraction: Double = 1.0) extends Product with Serializable

    Parameters to train the SVMWithSGD model

  16. class UncertaintySamplingRandomForest[C <: PointLabelingContext, G <: GroupLabelingContext] extends ActiveLearningAlgorithm[RandomForestModel, RandomForestParameters, C, G]

Value Members

  1. object DedupDemo

  2. object TreeDedupDemo

  3. object utils

    Utilities for the activeml package.