ML for CREs with Fermi-LAT

A cura di Nicolo' Cibrario, Raffaella Bonino (TO)

martedì 10 nov 2020, 15:00 → 17:00 Europe/Rome

Machine Learning for the measurement of the Cosmic-Ray Electron Spectrum with the Fermi Large Area Telescope

This presentation deals with the application of Machine Learning techniques to the computation of the Cosmic Ray Electrons energy spectrum with Fermi Large Area Telescope data.The most challenging part of the analysis concerns the event selection, i.e. the identification of electrons out of a predominant background of protons. 
Firstly I built an algorithm which exploits Neural Networks to identify electrons, using shower topology informations and drawing the outcome on the basis of MonteCarlo (MC) simulations. Secondly I developed algorithms based on clustering techniques, with the aim of unbinding the event selection from MC simulations, thus reducing the systematics of the measurement.
The performances achieved by the different models are compared and conclusions are discussed.

Presentazione_Seminario_10:11:20.pdf