Y. Meng, J. Huang, D. Chen, K Y. Hu, Y. Zhang, L M. Zhai, Y H. Zou, Y L. Yu, Y Y. Li
{"title":"Machine learning applications to energy reconstruction of gamma-ray showers for the Tibet AS\\(\\gamma \\) experiment","authors":"Y. Meng, J. Huang, D. Chen, K Y. Hu, Y. Zhang, L M. Zhai, Y H. Zou, Y L. Yu, Y Y. Li","doi":"10.1007/s10686-025-09993-9","DOIUrl":null,"url":null,"abstract":"<div><p>In order to improve the energy reconstruction accuracy of gamma-ray events observed by ground-based array experiments, this work propose a new energy estimator based on machine learning (ML) algorithm to determine the energies of gamma ray induced air showers in the energy range between 1 TeV and 10 PeV. We carry out a full Monte Carlo (MC) simulation using the Tibet air shower array and underground muon detector array, located at an altitude of 4,300 m above sea level. The MC simulated gamma-ray data are used to extract characteristic parameters depicting the air shower information, which are then fed into the ML model for training on both high-energy data sets (<span>\\(E >\\sim 10\\)</span> TeV) and low-energy data sets (<span>\\(E < 10\\)</span> TeV). In our simulation data tests, we found that the ML method showed significant advantages over traditional energy estimators (S50, <span>\\(N_e\\)</span>, and <span>\\(\\sum \\rho \\)</span>), with improved energy resolution for both low and high energy datasets. Compared to the traditional estimator, the energy resolution improves by approximately 30% for the inner array events and 55% for the outer array events at <span>\\(E < 10\\)</span> TeV. At around 100 TeV, the energy resolution for large zenith angle events in the outer array improves by approximately 20%. This work also found that while the energy resolution of events falling the inside array can only be slightly improved, however, events outside array and at large zenith shower clear improvements. Moreover, it is particularly noteworthy that the ML method has little difference in the energy resolution of the inner and outer array events. The enhanced energy resolution achieved through the machine learning method for outer array events reduces the limitations imposed by the observation area, resulting in an approximately 30% improvement in statistical events. This method is suitable for ground-based array experiments in gamma-ray astronomy, and provides some technical support for further study of the primary gamma-ray energy reconstruction.</p></div>","PeriodicalId":551,"journal":{"name":"Experimental Astronomy","volume":"59 2","pages":""},"PeriodicalIF":2.7000,"publicationDate":"2025-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s10686-025-09993-9.pdf","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Experimental Astronomy","FirstCategoryId":"101","ListUrlMain":"https://link.springer.com/article/10.1007/s10686-025-09993-9","RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ASTRONOMY & ASTROPHYSICS","Score":null,"Total":0}
引用次数: 0
Abstract
In order to improve the energy reconstruction accuracy of gamma-ray events observed by ground-based array experiments, this work propose a new energy estimator based on machine learning (ML) algorithm to determine the energies of gamma ray induced air showers in the energy range between 1 TeV and 10 PeV. We carry out a full Monte Carlo (MC) simulation using the Tibet air shower array and underground muon detector array, located at an altitude of 4,300 m above sea level. The MC simulated gamma-ray data are used to extract characteristic parameters depicting the air shower information, which are then fed into the ML model for training on both high-energy data sets (\(E >\sim 10\) TeV) and low-energy data sets (\(E < 10\) TeV). In our simulation data tests, we found that the ML method showed significant advantages over traditional energy estimators (S50, \(N_e\), and \(\sum \rho \)), with improved energy resolution for both low and high energy datasets. Compared to the traditional estimator, the energy resolution improves by approximately 30% for the inner array events and 55% for the outer array events at \(E < 10\) TeV. At around 100 TeV, the energy resolution for large zenith angle events in the outer array improves by approximately 20%. This work also found that while the energy resolution of events falling the inside array can only be slightly improved, however, events outside array and at large zenith shower clear improvements. Moreover, it is particularly noteworthy that the ML method has little difference in the energy resolution of the inner and outer array events. The enhanced energy resolution achieved through the machine learning method for outer array events reduces the limitations imposed by the observation area, resulting in an approximately 30% improvement in statistical events. This method is suitable for ground-based array experiments in gamma-ray astronomy, and provides some technical support for further study of the primary gamma-ray energy reconstruction.
为了提高地面阵列实验观测到的伽马射线事件的能量重建精度,本文提出了一种新的基于机器学习(ML)算法的能量估计器,用于确定1 TeV ~ 10 PeV能量范围内伽马射线诱导的空气簇射的能量。我们利用西藏空气淋点阵列和地下介子探测器阵列,在海拔4300米的高度进行了完整的蒙特卡罗(MC)模拟。MC模拟的伽马射线数据用于提取表征风淋信息的特征参数,然后将其输入ML模型,在高能数据集(\(E >\sim 10\) TeV)和低能数据集(\(E < 10\) TeV)上进行训练。在我们的模拟数据测试中,我们发现ML方法比传统的能量估计器(S50, \(N_e\)和\(\sum \rho \))具有显著的优势,并且在低能和高能数据集上都具有更高的能量分辨率。与传统估计器相比,能量分辨率提高了约30%% for the inner array events and 55% for the outer array events at \(E < 10\) TeV. At around 100 TeV, the energy resolution for large zenith angle events in the outer array improves by approximately 20%. This work also found that while the energy resolution of events falling the inside array can only be slightly improved, however, events outside array and at large zenith shower clear improvements. Moreover, it is particularly noteworthy that the ML method has little difference in the energy resolution of the inner and outer array events. The enhanced energy resolution achieved through the machine learning method for outer array events reduces the limitations imposed by the observation area, resulting in an approximately 30% improvement in statistical events. This method is suitable for ground-based array experiments in gamma-ray astronomy, and provides some technical support for further study of the primary gamma-ray energy reconstruction.
期刊介绍:
Many new instruments for observing astronomical objects at a variety of wavelengths have been and are continually being developed. Furthermore, a vast amount of effort is being put into the development of new techniques for data analysis in order to cope with great streams of data collected by these instruments.
Experimental Astronomy acts as a medium for the publication of papers of contemporary scientific interest on astrophysical instrumentation and methods necessary for the conduct of astronomy at all wavelength fields.
Experimental Astronomy publishes full-length articles, research letters and reviews on developments in detection techniques, instruments, and data analysis and image processing techniques. Occasional special issues are published, giving an in-depth presentation of the instrumentation and/or analysis connected with specific projects, such as satellite experiments or ground-based telescopes, or of specialized techniques.