Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
T: Fachverband Teilchenphysik
T 23: Experimentelle Techniken der Astroteilchenphysik 2
T 23.6: Vortrag
Montag, 27. März 2017, 18:00–18:15, S 055
Improvement of energy reconstruction by using machine learning algorithms in MAGIC — •Kazuma Ishio1, Galina Maneva2, Abelardo Moralejo3, David Paneque1, Julian Sitarek4, and Petar Temnikov2 for the MAGIC collaboration — 1Max-Planck-Institut für Physik, München, Germany — 2Institute for Nuclear Research and Nuclear Energy, Sofia, Bulgaria — 3Institut de Fisica d'Altes Energies (IFAE), Barcelona, Spain — 4University of Lodz, Lodz, Poland
The MAGIC telescopes perform gamma-ray astronomy at energies above 50 GeV and extending to about 50 TeV. The energy of the detected gamma ray is estimated with a set of parameters extracted from the shower image on the cameras and using Look-Up-Tables (LUTs) derived from Monte Carlo simulations. The current strategy yields an energy bias smaller than 5% with a resolution of approximately 20%, depending on energy range. The talk focuses on the usage of machine learning strategies, namely artificial neural network (ANN) and random forest (RF), for the determination of the gamma-ray energy. I will show that these strategies provide independent ways of reconstructing the energy, which are very helpful for cross-checks, and they also yield an improvement in the performance for energies above 1 TeV with respect to LUTs.