KernelDensity.jl

KernelDensity.jl copied to clipboard

Interface of KernelDensity.pdf is not consistent with Distributions.pdf

3

I've noticed `KernelDensity.pdf` cannot be used in the same way as `Distributions.pdf`. I think this is very confusing for users and is a potential source of errors. In particular, `pdf.(kde::UnivariateKDE,...

jaksle

Use AbstractFFTs instead of FFTW

6

The AbstractFFTs interface is a bit special: The only way to select a backend such as FFTW or RustFFT is to load the desired backend package (see e.g. https://github.com/JuliaMath/AbstractFFTs.jl/issues/32 and...

devmotion

https://github.com/JuliaStats/KernelDensity.jl/blob/ae70f461b6ac54c6d3d5e870deb140d21a8a8153/src/bivariate.jl#L21 Why do we carry this `permutedims`? This requires that `StatsPlots` explicitly perform the inverse transformation to plot a kde correctly, and it makes `Plots` default functions to plot the...

lmiq

allow Interpolations 0.15

1

KristofferC

Hey, love this package, and want to make sure I'm understanding how it works. For a bivariate kernel density, if you sum down a column, then multiply by the x...

EthanRLeonard

Multivariate implementation

6

This PR ports the previously existing bivariate implementation to a multivariate one, such that KernelDensity.jl now supports arbitrary dimensions. The bivariate estimation is now just a special case of the...

andreasKroepelin

Drop FFTW dependency

3

Since it effectively makes this package GPL licensed. Some possibilities might be - https://github.com/Taaitaaiger/RustFFT.jl (but https://github.com/Taaitaaiger/RustFFT.jl/issues/6) - https://github.com/dannys4/FFTA.jl (but https://github.com/dannys4/FFTA.jl/issues/46) cc @devmotion

andreasnoack