陈宁,毛善君,李德龙,岳俊

• 1:北京大学遥感与地理信息系统研究所
• 2:北京大学城市与环境学院

摘要(Abstract)：

针对通过数值方法对PM2.5进行预测已经取得了良好的效果,但相关模型重视时间影响因子而对空间影响因素的关联性考虑不足的问题,该文提出了多基站协同训练长短时记忆网络预测模型。该模型以时空数据作为输入,并将多个基站数据进行协同训练。MC-LSTM网络通过采用多基站共享参数的方式,减少了需要训练的网络复杂度,减轻了网络过拟合的风险。利用MC-LSTM网络对北京市21个监测基站数据进行了处理,结果表明:MC-LSTM网络能够同时对各个基站的PM2.5浓度进行预测。

关键词(KeyWords)： PM2.5预测;深度学习;LSTM模型;协同训练;空间因素

Abstract:

Keywords:

基金项目(Foundation): 国家重点研发计划项目(2016YFC0801800,2016YFC0801805);; 21世纪开放基金项目(21AT-2016-03)

作者(Author): 陈宁,毛善君,李德龙,岳俊

DOI: 10.16251/j.cnki.1009-2307.2018.07.014

参考文献(References)：

• [1]中华人民共和国环保部.2016中国环境状况公报[Z].北京:[出版者不详],[2017-05-31].(The Ministry of Environmental Protection China,China Environmental Status Bulletin 2016[Z].Beijing:[s.n.],[2017-05-31].)
• [2]杨新兴,冯丽华,尉鹏.大气颗粒物PM2.5及其危害[J].前沿科学,2012(1):22-31.(YANG Xinxing,FENG Lihua,WEI Peng.Air particulate matter PM2.5in Beijing and its harm[J].Frontier Science,2012(1):22-31.)
• [3]BYUN D W,CHING J K S,NOVAK J,et al.Development and implementation of the EPA’s Models-3initial operating version:community Multi-scale air quality(CMAQ)model[M]∥Air Pollution Modeling and Its Application XII.US:Springer,1998:357-368.
• [4]付倩娆.基于多元线性回归的雾霾预测方法研究[J].计算机科学,2016,43(s1):526-528.(FU Jingrao.Research on haze prediction based on multivariate linear regression[J].Compute Science,2016,43(s1):526-528.)
• [5]HU X,BELLE J H,MENG X,et al.Estimating PM2.5 concentrations in the conterminous United States using the random forest approach[J].Environmental Science&Technology,2017,51(12):6936.
• [6]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):1266-76.
• [7]LIU B,YAN S,LI J,et al.Forecasting PM2.5concentration using spatio-temporal extreme learning machine[C]∥IEEE International Conference on Machine Learning and Applications.[S.l.]:IEEE,2017:950-953.
• [8]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.)
• [9]范竣翔,李琦,朱亚杰,等.基于RNN的空气污染时空预报模型研究[J].测绘科学,2017,42(7):76-83.(FAN Junxiang,LI Qi,ZHU Yajie,et al.A Spatiotemporal prediction framework for air pollution based on deep RNN[J].Science of Surveying and Mapping,2017,42(7):76-83.)
• [10]ONG B T,SUGIURA K,ZETTSU K.Dynamic pretraining of deep recurrent neural networks for predicting environmental monitoring data[C]∥IEEE International Conference on Big Data.[S.l.]:IEEE,2014:760-765.
• [11]SNCHEZ A B,ORDEZ C,LASHERAS F S,et al.Forecasting SO2 pollution incidents by means of elman artificial neural networks and ARIMA models[J].Abstract and Applied Analysis,2013(3):1728-1749.
• [12]李小飞,张明军,王圣杰,等.中国空气污染指数变化特征及影响因素分析[J].环境科学,2012,33(6):1936-1943.(LI Xiaofei,ZHANG Mingjun,WANG Shengjie,et al.Variation characteristic and influencing factors of air pollution index in China[J].Environmental Science,2012,33(6):1936-1943.)
• [13]周景博,王鑫,杜婉君,等.北京市空气污染指数及其影响因素分析[J].中国环境监测,2015,31(2):53-56.(ZHOU Jingbo,WANG Xin,DU Wangjun,et al.Analysis of air pollution index and its influence factors in Beijing[J].Environmental Monitoring in China,2015,31(2):53-56.)
• [14]北京市环境保护监测中心[EB/OL].[2017-10-22].http:∥www.bjmemc.com.cn/g374/s1046/t1661.aspx.(Beijing Municipal Environmental Monitoring Center[EB/OL].[2017-10-22].http:∥www.bjmemc.com.cn/g374/s1046/t1661.aspx.)
• [15]刘鹏华,姚尧,梁昊,等.耦合卡尔曼滤波和多层次聚类的中国PM2.5时空分布分析[J].地球信息科学学报,2017,19(4):475-485.(LIU Penghua,YAO Yao,LIANG Hao,et al.Analyzing spatial-temporal distribution of PM2.5in China by integrating Kalman filter and Multi-level clustering[J].Journal of Geo-information Science,2017,19(4):475-485.)
• [16]CHEN Y,YANG K,JIE H,et al.Improving land surface temperature modeling for dry land of China[J].Journal of Geophysical Research Atmospheres,2011,116(D20):D20104.
• [17]何杰,阳坤.中国区域高时空分辨率地面气象要素驱动数据集[Z].寒区旱区科学数据中心,2011.(HE Jie,YANG Kun.China meteorological forcing dataset[Z].Cold and Arid Regions Science Data Center at Lanzhou,2011)
• [18]庄大方,刘纪远.中国土地利用程度的区域分异模型研究[J].自然资源学报,1997(2):105-111.(ZHUANG Dafang,LIU Jiyuan.Study on the model of regional differentiation of land use degree in China[J].Journal of Natural Resources,1997(2):105-111.)
• [19]HOCHREITER S,SCHMIDHUBER J.Long shortterm memory[J].Neural Computation,1997,9(8):1735-1780.
• [20]GOODFELLOW I,BENGIO Y,COURVILLE A.Deep Learning[M].[S.l.]:[s.n.],2016:404-407.
• [21]IOFFE S,SZEGEDY C.Batch Normalization:Accelerating Deep Network Training by Reducing Internal Covariate Shift[C]∥Intemational Conference on Machine Learning.[S.l.]:JMLR.org,2015:448-456.