Method for Separating Extensive Air Showers by Primary Mass Using Machine Learning for a Sphere-Type Cherenkov Telescope

Latypova, V.S.; Nemchenko, V.A.; Azra, C.G.; Bonvech, E.A.; Chernov, D.V.; Galkin, V.I.; Ivanov, V.A.; Podgrudkov, D.A.; Roganova, T.M.

Авторы: Latypova V.S., Nemchenko V.A., Azra C.G., Bonvech E.A., Chernov D.V., Galkin V.I., Ivanov V.A., Podgrudkov D.A., Roganova T.M.
Журнал: Moscow University Physics Bulletin
Том: 78
Номер: S1
Год издания: 2023
Издательство: Allerton Press
Местоположение издательства: New York, N.Y., United States
Первая страница: S25
Последняя страница: S31
DOI: 10.3103/s0027134923070196
Аннотация: The study of cosmic rays mass composition is an important problem in high-energy physics. The main goal of the SPHERE-2 experiment was to study the energy spectrum of the primary cosmic rays in the 10–300 PeV energy range. Also the experimental data allow approaching their mass composition. The separation of events into nuclei groups makes it possible to estimate the average masses over the sample. Using machine learning methods, we developed a separation method for the primary nuclei groups that formed extensive air showers based on the simulated events for the SPHERE-2 telescope. Various models of the high energy nucleus-nucleus interaction were used, but their predictions differ significantly. In the SPHERE-2 experiment data analysis, this problem was solved, first, by the use of the data on Cherenkov light, which has weak dependence on the model of hadronic interaction; second, the neural network was trained simultaneously on two interaction models (QGSJET-01 and QGSJETII-04), which differ greatly from each other. Therefore, the independence of experimental data processing from the choice of the nuclear interaction model was ensured. The regression task is solved by machine learning methods. The separation of events into three groups of nuclei—protons (p), nitrogen (N), and iron (Fe)—by using a neural network is more precise than that by using traditional methods.
Добавил в систему: Бонвеч Елена Алексеевна

	ИСТИНА	Войти в систему Регистрация
	ИСТИНА ИНХС РАН
	Главная Поиск Статистика О проекте Помощь

ИСТИНА

ИСТИНА ИНХС РАН

Method for Separating Extensive Air Showers by Primary Mass Using Machine Learning for a Sphere-Type Cherenkov Telescopeстатья Исследовательская статья

Прикрепленные файлы