Orateur
Description
This study presents a novel methodology for estimating uncertainty intervals in fits of Parton Distribution Functions (PDFs). By combining toy examples and fitting real-world PDFs, we critically evaluate the robustness of the Monte-Carlo (MC) replica method and the Hessian method in estimating credible intervals for PDFs. Our findings reveal that the methodologies typically used in PDF fits fail to replicate Bayesian interval estimates in some cases, mainly due to their inability to handle terms that are nonlinear as a function of the fitting parameters. We conclude by suggesting potential directions for further research to address these limitations and explore alternative approaches for constructing robust parameter interval estimates in PDF analyses.
