吴春霖,李琦,侯俊雄,KARIMIAN Hamed,陈工

• 1:北京大学遥感与地理信息系统研究所
• 2:北京大学智慧城市研究中心

摘要(Abstract)：

针对目前基于机器学习的PM2.5预报模型无法充分利用研究区域内其他相关站点的数据问题,该文提出了一种区域时空点数据的表示方法,并在此基础上提出了基于卷积神经网络的PM2.5预报模型。该模型利用了区域内多站点的历史PM2.5实测数据以及相应的气象预报数据,对区域内任一站点PM2.5浓度进行预报。实验结果显示,该模型在京津冀区域内能对未来至少3d内的PM2.5浓度进行较高精度的预报。与基于单站点的前馈神经网络预报结果对比表明,对区域整体污染及气象状况建模的卷积神经网络模型预报精度更高。该模型对区域内所有站点的预测结果与地面实测值的分布基本一致,表明了该模型具有对区域内PM2.5浓度进行时空预报的能力。

关键词(KeyWords)： 空气质量预报;细颗粒物PM2.5;卷积神经网络

Abstract:

Keywords:

基金项目(Foundation):

作者(Author): 吴春霖,李琦,侯俊雄,KARIMIAN Hamed,陈工

DOI: 10.16251/j.cnki.1009-2307.2018.08.011

参考文献(References)：

• [1]LEPEULE J,LADEN F,DOCKERY D,et al.Chronic exposure to fine particles and mortality:an extended follow-up of the Harvard Six Cities study from 1974to2009[J].Environmental Health Perspectives,2012,120(7):965.
• [2]QIU H,YU I T-S,WANG X,et al.Differential effects of fine and coarse particles on daily emergency cardiovascular hospitalizations in Hong Kong[J].Atmospheric Environment,2013,64(64):296-302.
• [3]ZHANG Y,BOCQUET M,MALLET V,et al.Realtime air quality forecasting,part I:History,techniques,and current status[J].Atmospheric Environment,2012,60(32):632-655.
• [4]STADLOBER E,HRMANN S,PFEILER B.Quality and performance of a PM10daily forecasting model[J].Atmospheric Environment,2008,42(6):1098-1109.
• [5]OSOWSKI S,GARANTY K.Forecasting of the daily meteorological pollution using wavelets and support vector machine[J].Engineering Applications of Artificial Intelligence,2007,20(6):745-755.
• [6]侯俊雄,李琦,朱亚杰,等.基于随机森林的PM2.5实时预报系统[J].测绘科学,2017,42(1):1-6.(HOU Junxiong,LI Qi,ZHU Yajie,et al.Real-time forecasting system of PM2.5concentration based on spark framework and random forest model[J].Science of Surveying and Mapping,2017,42(1):1-6.)
• [7]VOUKANTSIS D,KARATZAS K,KUKKONEN J,et al.Intercomparison of air quality data using principal component analysis,and forecasting of PM10 and PM2.5concentrations using artificial neural networks,in Thessaloniki and Helsinki[J].Science of the Total Environment,2011,409(7):1266-1276.
• [8]SUN W,ZHANG H,PALAZOGLU A,et al.Prediction of 24-hour-average PM2.5concentrations using a hidden Markov model with different emission distributions in Northern California[J].Science of the Total Environment,2013,443(3):93-103.
• [9]ONG B T,SUGIURA K,ZETTSU K.Dynamically pretrained deep recurrent neural networks using environmental monitoring data for predicting PM2.5[J].Neural Computing and Applications,2016,27(6):1553-1566.
• [10]KURT A,OKTAY A B.Forecasting air pollutant indicator levels with geographic models 3days in advance using neural networks[J].Expert Systems with Applications,2010,37(12):7986-7992.
• [11]FENG X,LI Q,ZHU Y,et al.Artificial neural networks forecasting of PM2.5pollution using air mass trajectory based geographic model and wavelet transformation[J].Atmospheric Environment,2015,107:118-128.
• [12]KRIZHEVSKY A,SUTSKEVER I,HINTON G E.Imagenet classification with deep convolutional neural networks[C]∥International Conference on Neural Information Processing Systems.[S.l.]:[s.n.],2012,60(2):1097-1105.
• [13]LESHNO M,LIN V Y,PINKUS A,et al.Multilayer feedforward networks with a nonpolynomial activation function can approximate any function[J].Neural Networks,1993,6(6):861-867.
• [14]CYBENKO G.Approximation by superpositions of a sigmoidal function[J].Mathematics of Control,Signals and Systems,1989,2(4):303-314.
• [15]HORNIK K,STINCHCOMBE M,WHITE H.Multilayer feedforward networks are universal approximators[J].Neural Networks,1989,2(5):359-366.
• [16]LECUN Y.Generalization and network design strategies[J].Connectionism in Perspective,1989:143-155.
• [17]SILVER D,HUANG A,MADDISON C J,et al.Mastering the game of go with deep neural networks and tree search[J].Nature,2016,529(7587):484-489.
• [18]ZHANG X,WANG Y,NIU T,et al.Atmospheric aerosol compositions in China:spatial/temporal variability,chemical signature,regional haze distribution and comparisons with global aerosols[J].Atmospheric Chemistry and Physics,2012,12(2):779-799.