Package io.github.ai4ci.abm.riskmodel

Interface Kernel

All Superinterfaces:
Serializable
All Known Implementing Classes:
ImmutableKernel
@Immutable public interface Kernel extends Serializable
A convolution kernel or filter

  • Method Summary

    Modifier and Type
    Method
    Description
    static Kernel
    empty()
     
    static Kernel
    from(int offset, double... values)
     
    double[]
    getDensity()
    Kernel values, offset as required, compared to a convolution this is in reverse order.
    default double
    getDensity(int offset)
    get the kernel density value relevant to a point offset from now positive values of the offset are in the past, negative values relate to future data the full range of data is from [ -prospectiveSize(),retrospectiveSize() )
    int
    getOffset()
    index of zero point in values. positive values mean that the kernel is 2 sided. zero means only left sided. negatives don't make particular sense but should work consistently
    default Kernel
    normaliseTo(double sum)
     
    default int
    prospectiveSize()
    how far does this kernel extend in the future (inclusive)?
    default int
    retrospectiveSize()
    how far in the past does this kernel extend (exclusive)?
    default Kernel
    scale(double factor)
     
    default int
    size()
     
    static Kernel
    square(int offset, int length)
     

  • Method Details

    • getDensity

      double[] getDensity()
      Kernel values, offset as required, compared to a convolution this is in reverse order. The first value in this array is applied to the most recent data point, which may be in the future (prospective).

    • getOffset

      int getOffset()
      index of zero point in values. positive values mean that the kernel is 2 sided. zero means only left sided. negatives don't make particular sense but should work consistently

    • getDensity

      default double getDensity(int offset)
      get the kernel density value relevant to a point offset from now positive values of the offset are in the past, negative values relate to future data the full range of data is from [ -prospectiveSize(),retrospectiveSize() )

    • prospectiveSize

      default int prospectiveSize()
      how far does this kernel extend in the future (inclusive)?

    • retrospectiveSize

      default int retrospectiveSize()
      how far in the past does this kernel extend (exclusive)?

    • size

      default int size()

    • scale

      default Kernel scale(double factor)

    • normaliseTo

      default Kernel normaliseTo(double sum)

    • from

      static Kernel from(int offset, double... values)
      Parameters:
      offset - : index of zero point in values. positive values mean that the kernel is 2 sided. zero means only left sided. negatives don't make particular sense but should work consistently
      values -
      Returns:

    • empty

      static Kernel empty()

    • square

      static Kernel square(int offset, int length)