Hi,
eg. for pdf/K-factor system error, they are correlated for different bins.
How to deal with that?
Does TLimit support the function ?
Thanks
Yi
Hi,
eg. for pdf/K-factor system error, they are correlated for different bins.
How to deal with that?
Does TLimit support the function ?
Thanks
Yi
Hi,
I don’t think TLimit support the inclusion of systematic errors with correlation between bins. It is able only to perform uncorrelated fluctuations of the bin content.
It is maybe possible to model these correlation using RooFit, this should be checked, and then use it as input to the HybridCalculator class of RooStats
which uses the same statistical technique of TLimit.
See root.cern.ch/root/htmldoc/RooSta … lator.html
Best Regards
Lorenzo
Thanks.
[quote=“moneta”]Hi,
I don’t think TLimit support the inclusion of systematic errors with correlation between bins. It is able only to perform uncorrelated fluctuations of the bin content.
It is maybe possible to model these correlation using RooFit, this should be checked, and then use it as input to the HybridCalculator class of RooStats
which uses the same statistical technique of TLimit.
See root.cern.ch/root/htmldoc/RooSta … lator.html
Best Regards
Lorenzo[/quote]