KernelDensity with range

[juliastats]

KernelDensity with range

[simonbyrne]

I would like to have a range and guarantee that the function is 0 (dies down) at the boundaires. How can I setup KernelDensity properly to achieve this result?

Short answer: you can't. The problem is that the convolution in KDEs is based on a Fourier transform, which essentially assumes that your domain is periodic: this means you need to pad the extent of each range so that this "wrap-around" effect is small.

What you could do is construct the kde, trim off what you don't want, and put it back together, e.g.:

X = rand(10)
r = linrange(-1.0,2.0,1024)
# construct "padded" kde
k = kde(X,r)

# lower and upper indices
i,j = searchsortedfirst(r,0), searchsortedlast(r,1)
# construct trimmed kde object
k2 = UnivariateKDE(k.x[i:j],k.density[i:j])

[skrisna]

That is great. Thanks for the exposition and the padding suggestion.