QPSO-BP神经网络算法解析地质样品组分  

BP Neural Network Algorithms of Quantum Particle Swarm Optimization for Geological Sample Composition Analysis

作　　者：常兰[1,2] 周洪祥 蔡小杰 颜瑜成 周益民 CHANG Lan;ZHOU Hong-xiang;CAI Xiao-jie;YAN Yu-cheng;ZHOU Yi-min(Chengdu University of Technology College of Engineering Technology,Leshan Sichuan 614000,China;Southwest Institute of Physics,Chengdu 610225,China)

机构地区：[1]成都理工大学工程技术学院,四川乐山614000 [2]核工业西南物理研究院,成都610225

出　　处：《核电子学与探测技术》2019年第4期523-528,共6页Nuclear Electronics & Detection Technology

基　　金：四川省教育厅项目(18ZA0065)资助

摘　　要：BP神经网络和量子粒子群算法目前被广泛应用于众多领域中。通过运用Matlab对地质样品成分的分析,将BP神经网络与基于量子粒子群算法的BP神经网络的性能作对比,在对地质样品的X荧光数据分析处理过程中,QPSO-BP神经网络预测精度明显高于传统BP模型,稳定性更强。BP neural network and quantum particle swarm optimization are widely used in many fields.By using Matlab to analyze the composition of geological samples,the performance of BP neural network is compared with that of BP neural network based on quantum particle swarm optimization.In the process of X-ray fluorescence data analysis and processing of geological samples,the prediction accuracy of QPSO-BP neural network is obviously higher than that of traditional BP model,and its stability is stronger.

关 键 词：BP神经网络 量子粒子群算法 地质样品成分分析 QPSO-BP神经网络算法 

分 类 号：TL99[核科学技术—核技术及应用]