KernelDensityEstimation

public class KernelDensityEstimation

Kernel density estimation using a Gaussian kernel and automatic bandwidth selection via Silverman’s rule-of-thumb.

  • Bandwidth used in evaluate() method

    Declaration

    Swift

    public let bandwidth: Double

  • Creates a new instance from an array of Doubles

    Declaration

    Swift

    public init?(_ data: [Double], bandwidth: Double? = nil)

    Parameters

    data

    Array of Doubles. If automatic bandwidth selection is required, this arry needs to have no fewer than two values.
    bandwidth

    Either specify the size of the bandwidth (the standard deviation of the normal kernel), or specify nil to automatically select the bandwidth using Silverman’s rule-of-thumb. Note that this rule is only appropriate if the data comes from a distribution not too dissimilar to a normal distribution. In other words, you will be better served by manually specifying the bandwidth if the source distribution is multi-modal. See the Kernel Density Estimation Wikipedia article.

    Return Value

    A new instance of KernelDensityEstimation, or nil if data contained fewer than two values and bandwidth was set to nil.

  • Evaluates the Kernel Density Estimator at the given value, x.

    Declaration

    Swift

    public func evaluate(_ x: Double) -> Double

    Parameters

    x

    The point at which the KDE should be evaluated.

    Return Value

    The density at the specified point, x.