Kernel Functions
A kernel function defines the smoothing behavior of the ASH estimator.
AverageShiftedHistograms can use any function as a kernel that satisifies the following properties:
- Non-negative:
f(x) >= 0 - Symmetric:
f(x) = f(-x)
In most applications, kernels need to integrate to 1, but that is not required in AverageShiftedHistograms
Function in AverageShiftedHistograms.Kernels
The Kernels module defines a collection of kernels:
biweight(default)cosineepanechnikovtriangulartricubetriweightuniformgaussianlogistic
See https://en.wikipedia.org/wiki/Kernel(statistics)#Kernelfunctionsincommon_use for detailed descriptions.
