TKDE or TH1K for weighted events

Roofit and RooStats
1

Hi,

I’m trying to approximate some distributions that are used as a template for a fit. The distributions are a bit spiky due to a lack of statistics and the hope was, that a reasonable approximation would be smoother and therefore easier to fit.
The first things I found were the TKDE and TH1K classes, but both do not support weighted entries. My templates are generated from a Monte Carlo data set with weighted entries. Is there a generalization of one of these classes for weighted entries, or even a better way to do something like this?

Thank you,
Jan

2

Hi,

It is true TKDE does not support weighted events. It is something that needs to be implemented in the class.
However, you can try to use then the class RooKeysPdf of Roofit, I think it supports weighted events
(see root.cern.ch/root/html/RooKeysPdf.html )

Best Regards

Lorenzo