classification

Type Members

1. trait BinaryClassificationModel extends MultinomialClassificationModel

   

   Trait for binary classifiers.

   Trait for binary classifiers.

   This is only used for binary classifiers. See MultinomialClassificationModel for multinomial classifiers.

2. trait ClassificationModel extends AnyRef

   

   Trait for all classification models.

3. case class DecisionTreeClassifierModel(rootNode: Node, numFeatures: Int, numClasses: Int) extends MultinomialClassificationModel with DecisionTree with Serializable with Product

   

   Class for decision tree classification models.

   Class for decision tree classification models.

   rootNode

   root decision tree node

   numFeatures

   number of features used in prediction

   numClasses

   number of predictable classes

4. case class GBTClassifierModel(trees: Seq[DecisionTreeRegressionModel], treeWeights: Seq[Double], numFeatures: Int) extends BinaryClassificationModel with TreeEnsemble with Serializable with Product

   

   Class for a gradient boost classifier model.

   Class for a gradient boost classifier model.

   trees

   trees in the gradient boost model

   treeWeights

   weights of each tree

   numFeatures

   number of features

5. case class LogisticRegressionModel(coefficients: Vector, intercept: Double, threshold: Option[Double] = Some(0.5)) extends BinaryClassificationModel with Serializable with Product

   

   Class for binary logistic regression models.

   Class for binary logistic regression models.

   coefficients

   coefficients vector for model

   intercept

   intercept of model

   threshold

   threshold for pegging predictions

6. trait MultinomialClassificationModel extends ClassificationModel

   

   Trait for classification models.

   Trait for classification models.

   This trait handles multinomial classification models as well as binary classification models.

   Annotations

   @SparkCode()

7. case class OneVsRestModel(classifiers: Array[BinaryClassificationModel]) extends Product with Serializable

   

   Class for multinomial one vs rest models.

   Class for multinomial one vs rest models.

   One vs rest models are comprised of a series of BinaryClassificationModels which are used to predict each class.

   classifiers

   binary classification models

8. case class RandomForestClassifierModel(trees: Seq[DecisionTreeClassifierModel], treeWeights: Seq[Double], numFeatures: Int, numClasses: Int) extends MultinomialClassificationModel with TreeEnsemble with Serializable with Product

   

   Class for random forest classification models.

   Class for random forest classification models.

   trees

   trees of the random forest

   numFeatures

   number of features in feature vector

   numClasses

   number of predictable classes

9. case class SupportVectorMachineModel(coefficients: Vector, intercept: Double, threshold: Option[Double] = ...) extends BinaryClassificationModel with Serializable with Product

   

   Class for support vector machine models.

   Class for support vector machine models.

   coefficients

   coefficients of SVM

   intercept

   intercept for SVM

   threshold

   threshold for pegging prediction