LNL: P. NAPIRALLA, BAYESIAN METHODS IN NUCLEAR STRUCTURE PHYSICS
LABORATORI NAZIONALI DI LEGNARO
The principle of Bayesian inference is used in many different fields of science, e.g. medicine and computer science. The foundation of Bayesian inference lies in Bayes' theorem, which offers a powerful alternative method for data analysis. Nevertheless, Bayesian inference is still rather unpopular in fields like nuclear structure physics, where very sensitive detector systems are needed. In g-spectroscopy, one of the essential experimental tools of nuclear structure physics, the state-of-the-art detector systems are highly segmented High-Purity Germanium detectors like the Advanced GAmma Tracking Array AGATA. Due to AGATA's Germanium shell without any Compton-shielding, g-ray tracking algorithms are needed. The mathematical problem these g-ray tracking algorithms are based on, forms a perfect example case for the benefits of Bayesian inference over standard statistical inference methods. Using basic terms of probability theory, a short introduction into Bayesian inference is given and essential principles are presented. In addition, a how-to approach of Bayesian inference to the principle of g-ray tracking is shown in the form of the Fuzzy Bayes Tracking algorithm. Possible difficulties, as well as benefits of Bayesian inference are elaborated in detail.

DATA: 22-03-2018

Sito Collegato : https://agenda.infn.it/conferenceDisplay.py?confId=15382

Istituto Nazionale di Fisica Nucleare - Piazza dei Caprettari, 70 - 00186 Roma
tel. +39 066840031 - fax +39 0668307924 - email: presidenza@presid.infn.it