-
基于热中子探测器实验模拟数据, 使用决策树(decision tree, DT)、随机森林(random forest, RF)和BP神经网络(back-propagation neural network, BPNN)构建了宇宙线粒子鉴别机器学习模型, 对每种粒子分别使用不同的机器学习算法基于模拟数据进行模型训练, 并针对算法进行超参数调整, 将每种算法的AUC值和 Q品质因子作为粒子成分鉴别的评价指标. 实验结果表明, 不同机器学习模型对粒子预测精度影响很大. 在测试检验中, 经过交叉网格搜索方法调参后的决策树鉴别模型对中成分(碳氮氧和镁铝硅)比较敏感, 鉴别模型AUC值均在0.95以上, Q品质因子均大于6; 经交叉网格搜索方法调参后的随机森林鉴别模型对于宇宙线粒子鉴别的效果最好, 所有粒子鉴别模型的AUC值均大于0.92且 Q品质因子均在4以上; BP神经网络算法只对质子和铁核比较敏感. 本研究对宇宙线粒子鉴别和筛选提供了新的方法和选择, 可为热中子探测器后续开展宇宙线能谱测量提供新思路.Machine learning algorithms can learn the rules and patterns of big data through computers, excavate potential information hidden behind the data, and be widely used to solve classification, regression, clustering, and other problems. Firstly, this paper uses CORSIKA software to simulate the process of cosmic ray cascade shower in the atmosphere, generating information such as the initial energy, zenith angle, azimuth angle of cosmic ray particles. Then, this paper uses the Geant4 toolkit to conduct thermal neutron detector response simulation, generating 4000 particles in each of proton, helium, CNO, MgAlSi and iron. Based on the experimental simulation data of thermal neutron detector, this paper constructs machine learning models for identifying cosmic ray particles by using decision tree (DT), random forest (RF) and BP neural network (BP NN) respectively. For each particle, all the machine learning algorithms are used for model training based on the simulation data. The cross grid search method is used to adjust the hyper parameters of each machine learning algorithm. The AUC value and Qquality factor value of each algorithm are used as evaluation indexes for particle composition identification. The AUC value is a general indicator for evaluating algorithm performance in machine learning and the Qquality factor value is an evaluation index commonly used in the field of high energy physics. The Experimental results show that different machine learning models have great influence on particle prediction accuracy, and the random forest cosmic ray particle identification model has sufficient accuracy and generalization capability. In the test, the decision tree algorithm adjusted by cross grid search method is sensitive to the medium components (CNO and MgAlSi). The AUC values of the algorithm are all above 0.95 and the Qquality factor values are all above 6. The random forest algorithm adjusted by the cross grid search method has the best effect on the identification of cosmic ray particles. The AUC values of the algorithm are all more than 0.92 and the Qquality factor values are all more than 4. The BP neural network algorithm is only sensitive to proton and iron. This study provides a new method and selection for identifying and screening the cosmic ray particles and it also provides a new idea for the following measurement of cosmic ray energy spectrum by thermal neutron detector.
-
Keywords:
- cosmic rays/
- particle identification/
- machine learning/
- random forest
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] -
超参数 目标成分 质子 氦核 碳氮氧 镁铝硅 铁核 criterion Entropy Entropy Entropy Entropy Entropy max_depth 21 29 40 28 19 min_samples_split 2 4 7 2 4 min_weight_fraction_leaf 0 0 0 0 0 min_samples_leaf 1 1 1 1 1 
下载:
导出CSV
超参数 目标成分 质子 氦核 碳氮氧 镁铝硅 铁核 criterion Gini Gini Entropy Entropy Entropy n_estimators 48 88 30 15 21 max_depth 20 26 30 27 23 min_samples_split 2 2 2 1 2 min_samples_leaf 1 1 1 1 1 下载:
导出CSV
训练结果 隐藏节点个数 5 6 7 8 9 10 11 12 13 迭代次数 20000 20000 20000 25000 27000 20000 20000 20000 20000 算法AUC值 0.5503 0.5045 0.5293 0.5593 0.6329 0.6276 0.6177 0.6142 0.6418 Q品质因子 0.82 0.29 0.58 0.86 1.26 1.25 1.22 1.26 1.34 下载:
导出CSV
超参数 目标成分 质子 氦核 碳氮氧 镁铝硅 铁核 隐藏层节点数 13 11 13 13 11 初始学习率 0.01 0.01 0.01 0.01 0.01 迭代次数 20000 25000 20000 20000 20000 下载:
导出CSV
目标成分 效率/% 纯度/% BP神经网络 决策树 随机森林 BP神经网络 决策树 随机森林 质子 64.9 74.8 75.7 74.4 77.6 91.1 氦核 36.0 83.3 79.3 52.8 80.1 95.7 碳氮氧 10.3 93.4 81.5 64.5 94.8 99.4 镁铝硅 16.9 91.8 78.7 69.9 92.1 95.8 铁核 82.8 88.1 91.1 87.5 88.7 93.5 下载:
导出CSV
目标成分 AUC Q品质因子 BP神经网络 决策树 随机森林 BP神经网络 决策树 随机森林 质子 0.7962 0.8555 0.9247 2.71 3.15 5.42 氦核 0.6418 0.8805 0.9537 1.34 3.75 8.38 碳氮氧 0.5444 0.9612 0.9739 0.87 7.55 20.1 镁铝硅 0.5754 0.9504 0.9531 1.25 6.54 8.39 铁核 0.8751 0.8952 0.9380 2.96 2.97 4.40 下载:
导出CSV
-
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]
下载:
计量
- 文章访问数:3037
- PDF下载量:89
- 被引次数:0