SEMINAR

A multidimensional unfolding method based on Bayes' theorem

Giulio D'Agostini

(University of Rome La Sapienza)

Sala PA1, IST, Edifício Matemática
Friday, April 17th, 2009 at 04:00 PM

Abstract

Probability inversion (hence "Bayesian") is applied to unfold measured distributions in order to correct them for distortions due instrumental and physical effects, that might also include background. The ideas behind the program unfolding, widely used in particle and astroparticle physics, are presented, together with improvements to better handle small numbers and the uncertainty associated to the resulting spectrum.