摘要
作为大气污染主要因素之一的可吸入颗粒物PM_(2.5)严重影响人类的健康,受到广泛的关注。科学高效地预测PM_(2.5)有利于人类提前做好防护措施,保护自身安全。为此设计了基于时域卷积神经网络的PM_(2.5)浓度预测模型,选取中国环境监测总站的全国城市空气质量实时发布平台的数据,对陕西省西安市的PM_(2.5)浓度进行了预测,并对预测结果进行分析。与长短时记忆神经网络和门控循环单元模型进行对比实验,结果表明时域卷积神经网络在预测PM_(2.5)浓度中具有较好的性能。
As one of the main factors of air pollution,PM_(2.5)seriously affects human health,and attracts increasing attention of human beings.Scientific and efficient prediction of PM_(2.5)is conducive to human protection measures in advance to protect their own safety.In this paper,a prediction model of PM_(2.5)concentration based on time-domain convolutional neural network is designed,and the PM_(2.5)concentration in Xi’an city of Shaanxi province is predicted by selecting data from the real-time release platform of National Urban Air Quality of China Environmental Monitoring Station,and the prediction results are analyzed.Compared with the long and short memory neural network and gated cyclic unit model,the results show that the convolutional neural network has better performance in predicting PM_(2.5)concentration.
作者
任瑛
马乐荣
夏必胜
REN Ying;MA Lerong;XIA Bisheng(College of Mathematics and Computer Science,Yan’an University,Yan’an 716000,China;Yan’an Key Laboratory of Intelligent Information Processing of Big Data in Red Culture,Yan’an 716000,China)
出处
《微型电脑应用》
2024年第4期89-92,共4页
Microcomputer Applications
基金
延安大学十四五中长期重大科研项目(2021ZCQ012)
延安大学产学研合作培育项目(CXY202107)。
