2D KDE on root histogram

[Alex_Marshall]

Hello All

I have a 2D histogram in root which I need to fit a KDE to, which I will in future sample from. I need to keep the data binned as the number of entries is high and a KDE process looking at every point would take too long.

Anyone know the best way to approach this?

Thanks a lot

Alex

[ksmith]

This is the status as I currently understand it:

Two dimensional KDE

Hi, Sorry for the delay. We have in RooFit a n-dimensional kernel density estimator class, RooNDKeysPdf. The class is not very efficient, especially if you have a large amount of data. We are planning to have a new one based on TKDTree, but so far we did not have the resources to develop it. We are planning to allocate a student to work on this during the next summer. We will let you know as soon we have something ready to be used Best Regards Lorenzo

Maybe @moneta has an update.

[moneta]

HI,

This is correct, for multi-dimension the only implementation that we have currently is the one in RooFit but it is not very efficient.
As mentioned before, there has been some development for a new KDE class for multi-dimension. This work has not been merged yet in ROOT, but a development version is available in GitHub at this link,

vukasinmilosevic/TKNDTree

You are welcomed to try it and give us eventually feedback.

Lorenzo