Speaker
Dr
Andy Buckley
(University of Glasgow)
Description
We propose a robust way of estimating and propagating uncertainties in the "tuning" of Monte Carlo event generator simulations to describe experimental data. This approach enhances the established eigentunes method by using a statistically sound coverage measure rather than an ad hoc tolerance scale to determine the size of parameter variations allowed by the data. The efficacy of the new scheme, and wider reflections on the nature of tuning and MC uncertainties, will be illustrated by application of the Professor tool to toy and real tuning problems.
Summary
We propose a robust way of estimating and propagating uncertainties in the "tuning" of Monte Carlo event generator simulations to experimental data, implemented in the Professor tool.
Primary author
Dr
Andy Buckley
(University of Glasgow)
Co-author
Holger Schulz