AdaBoostClassifier

public class AdaBoostClassifier

Undocumented

  • decision stumps for the booster

    Declaration

    Swift

    public var stumps: [DecisionTree]

  • weights computed during boosting

    Declaration

    Swift

    public var weights: [Float]

  • alphas stored during boosting

    Declaration

    Swift

    public var alphas: [Float]

  • errors used during boosting

    Declaration

    Swift

    public var errors: [Float]

  • number of stumps to be created

    Declaration

    Swift

    public var iterations: Int

  • training data

    Declaration

    Swift

    public var data: [[String]]
  • K

    K value for boosting

    Declaration

    Swift

    public var K: Float

  • column number of target var

    Declaration

    Swift

    public var target: Int

  • choice of regression vs. classification

    Declaration

    Swift

    public var using: String

  • initializer for AdaBoost Classifier

    Declaration

    Swift

    public init(data: [[String]], target: Int, till: Int, using: String)

    Parameters

    data

    data with labels
    target

    column number of the labels
    till

    number of stumps to be generated
    using

    info gain or gini Impurity

  • Creates the stumps and fills weights

    Declaration

    Swift

    public func boost()

    Return Value

    None

  • Computes Error for a given stump

    Declaration

    Swift

    public func computeError(stump: DecisionTree)

    Parameters

    stump

    Decision Tree for which error needs to be computed

    Return Value

    Nothing

  • Returns new dataset with wieghted sampling based on previous errors

    Declaration

    Swift

    public func getWeightSampledData(from: [[String]]) -> [[String]]

    Parameters

    from

    data to sample from

    Return Value

    sampled data

  • Classfies an example by getting all stump classification and weighing them

    Declaration

    Swift

    public func classify(this: [[String]]) -> String

    Return Value

    classification as a stringß

  • Scores the booster’s accuracy on test data

    Declaration

    Swift

    public func score(testData: [[String]]) -> (Float, [String])

    Parameters

    testData

    test data as a 2D string array with feature header

    Return Value

    • accuracy of classifications as float
    • classifications as string array