[1]赵文芳,林润生,唐 伟,等.基于深度学习的PM2.5短期预测模型[J].南京师范大学学报(自然科学版),2019,42(03):32-41.[doi:10.3969/j.issn.1001-4616.2019.03.005]
　Zhao Wenfang,Lin Runsheng,Tang Wei,et al.Forecasting Model of Short-Term PM2.5 ConcentrationBased on Deep Learning[J].Journal of Nanjing Normal University(Natural Science Edition),2019,42(03):32-41.[doi:10.3969/j.issn.1001-4616.2019.03.005]
点击复制

基于深度学习的PM2.5短期预测模型()
分享到：

《南京师范大学学报》（自然科学版）[ISSN:1001-4616/CN:32-1239/N]

卷:
第42卷
期数:
2019年03期
页码:
32-41
栏目:
·全国机器学习会议论文专栏·
出版日期:
2019-09-30

文章信息/Info

Title:
Forecasting Model of Short-Term PM2.5 ConcentrationBased on Deep Learning
文章编号:
1001-4616(2019)03-0032-10
作者:
赵文芳1; 2; 林润生2; 唐 伟3; 周 勇3
(1.北京城市气象研究院,北京 100089)(2.北京市气象信息中心,北京 100089)(3.中国气象局发展研究中心,北京 100081)
Author(s):
Zhao Wenfang1; 2; Lin Runsheng2; Tang Wei3; Zhou Yong3
(1.Institute of Urban Meteorology,China Meteorological Administration,Beijing 100089,China)(2.Beijing Meteorological Information Center,Beijing Meteorological Bureau,Beijing 100089,China)(3.Development and Research Center,China Meteorological Administration,Beijing 100081,China)
关键词:
PM2.5浓度预测; 机器学习; 长短时记忆; 深度学习; 卷积神经网络
Keywords:
forecast for PM2.5 concentration; machine learning; long short term memory; deep learning; convolutional neural network
分类号:
TP391
DOI:
10.3969/j.issn.1001-4616.2019.03.005
文献标志码:
A
摘要:
为了提高PM2.5浓度短期预报的准确率,解决现有PM2.5浓度短期预报准确率不高的问题,提出了一种基于卷积神经网络和长短时记忆的深度学习预报测型. 首先,综合考虑气温、相对湿度、降水量、风力、能见度等多种气象要素,综合分析气象要素与PM2.5浓度相关性. 其次,利用PM2.5浓度数据、气象站点观测数据和气象要素网格实况分析数据进行融合处理,生成用于训练和测试的时空序列数据,并使用卷积神经网络和长短时记忆网络获取时空特征. 通过大量实验确定模型中关键参数,然后利用最优参数建立预测模型. 最后,使用模型对PM2.5未来24 h浓度进行预测,并与支持向量机、业务中的预报模型进行对比. 实验结果表明,相比其他机器学习方法和预报方法,卷积神经网络和长短时记忆相结合的预测方法能有效提高PM2.5浓度未来24 h预测精度,并具有较高的泛化能力.
Abstract:
In order to improve the accuracy of PM2.5 concentration forecast in Beijing Meteorological Bureau,a deep learning prediction model based on convolutional neural network(CNN)and long short term memory neural network(LSTM)was proposed. Firstly,the feature vectors extraction was carried out by using the correlation analysis technique from meteorological data such as temperature,wind,relative humidity,precipitation,visibility and atmospheric pressure. Secondly,taking into account the fact that PM2.5 concentration was significantly affected by surrounding meteorological impact factors,meteorological grid analysis data were novel involved into the model,as well as the historical PM2.5 concentration data and meteorological observation data of the present station. Spatiotemporal sequence data were generated from these data after integrated processing. Highlevel spatiotemporal features were extracted through the combination of the CNN and LSTM. Finally,future 24-hour prediction of PM2.5 concentration was made by the model. The comparison among the accuracy of this optimized model,support vector machine(SVM)and existing PM2.5 forecast system is performed to evaluate their performance. The results show that the proposed CNN-LSTM model performs better than SVM and current operational models in Beijing Meteorological Bureau,which has effectively improved the prediction accuracy of PM2.5 concentration for different time predictions scales in the next 24 hours.

参考文献/References:

[1] KAUFMAN Y J,DIDIER T,OLIVIER B. A satellite view of aerosols in the climate system[J]. Nature,2002,419:215-233.
[2]CHAN C,YAO X. Air pollution in mega cities in China[J]. Atmospheric environment,2008,42(1):1-12.
[3]张西雅,扈海波. 基于多源数据的北京地区PM2.5暴露风险评估[J]. 北京大学学报(自然科学版),2018,54(5):1103-1113.
[4]JIN Q,FANG Y,WEN B,et al. Spatio-temporal variations of pm2.5 emission in china from 2005 to 2014[J]. Chemosphere,2014,183:429-436.
[5]Van DONKELAAR A,MARTIN R V,BRAUER M,et al. Use of satellite observations for long-term exposure assessment of global concentrations of fine particulate[J]. Environmental health perspectives,2015,123(2):135-143.
[6]FANG C L,WANG Z B,XU G. Spatial-temporal characteristics of pm2.5 in china:A city level perspective analysis[J]. Journal of geographical sciences,2016,26:1519-1532.
[7]LIM S S,VOS T,FLAXMAN A D,et al. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions,1990-2010:a systematic analysis for the Global Burden of Disease Study 2010[J]. The lancet,2012,380:2224-2260.
[8]CHEN R,WANG X,MENG X,et al. Communicating air pollution-related health risks to the public:an application of the air quality health index in Shanghai,China[J]. Environment international,2013,1(5):168-173.
[9]CHEN J,LU J,AVISE J C,et al. Seasonal modeling of PM 2.5 in California’s San Joaquin Valley[J]. Atmospheric environment,2014,92:182-190.
[10]WU Q Z,SHI A,LI Y,et a1. Air quality forecast of PMl0 in Beijing with community multi-scale air quality modeling(CMAQ)system:emission and improvement[J]. Geoscientific model development,2014,7(5):2243-2259.
[11]邓涛,吴兑,邓雪娇,等. 珠三角空气质量暨光化学烟雾数值预报系统[J]. 环境科学与技术,2013(4):62-68.
[12]周广强,谢英,吴剑斌,等. 基于WRF-Chem模式的华东区域PM2.5预报及偏差原因[J]. 中国环境科学,2016,36(8):2251-2259.
[13]高怡,张美根. 2013年1月华北地区重雾霾过程及其成因的模拟分析[J]. 气候与环境研究,2014,19(2):140-152.
[14]张恒德,咸云浩,谢永华,等. 基于时间序列分析和卡尔曼滤波的霾预报技术[J]. 计算机应用,2017,37(11):279-284.
[15]赵秀娟,徐敬,张自银,等. 北京区域环境气象数值预报系统及PM2.5预报检验[J]. 应用气象学报,2016,27(2):160-172.
[16]WANG T J,JIANG F,DENG J J,et al. Urban air quality and regional haze weather forecast for Yangtze River Delta[J]. Atmospheric environment,2012,58(15):70-83.
[17]刘杰,杨鹏,吕文生,等. 模糊时序与支持向量机建模相结合的PM2.5质量浓度预测[J]. 北京科技大学学报,2014,36(12):1694-1703.
[18]李龙,马磊,贺建峰,等. 基于特征向量的最小二乘支持向量机PM2.5浓度预测模型[J]. 计算机应用,2014,34(8):2212-2216.
[19]VOUKANTSIS D,KARATZAS K,KUKKONEN J,et al. Inter-comparison of air quality data using principal component analysis,and forecasting of PM10 and PM2.5 concentrations using artificial neural networks,in Thessaloniki and Helsinki[J]. Science of the total environment,2011,409(7):1266-1276.
[20]MISHRA D,GOLYAL P,UPADHYAY A. Artificial intelligence based approach to fore-cast PM2.5 during haze episodes:a case study of Delhi,India[J]. Atmospheric environment,2015,120:239-248.
[21]LI J D,CHANG J Z,LEI M M. Dynamic forecasting model of short-term PM2.5 concentration based on machine learning[J]. Journal of computer applications,2017,37(11):3057-3063.
[22]JIA C Z,FANG D,YE Y C,et al. Long short-term memory-fully connected(LSTM-FC)neural network for PM2.5 concentration prediction[J]. Chemosphere,2019,220:486-492.
[23]ZHENG Y,YI X,LI M,et al. Forecasting fine-grained air quality based on big data[C]//Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Sydney,2015.
[24]YI X,ZHENG Y,ZHANG J,et al. ST-MVL:filling missing values in geo-sensory time series data[C]//Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence. Palo Alto:AAAI Press,2016.
[25]VIDUSHI C,ANAND D,VIJAYANAND K,et al. Time Series Based LSTM Model to Predict Air Pollutant’s Concentration for Prominent Cities in India[C]//Proceedings of the first International Workshop on Utility-Driven Mining. London,2018.
[26]WEBER S A,INSAF T Z,HALL E S,et al. Assessing the impact of fine particulate matter(PM2.5)on respiratory-cardiovascular chronic diseases in the New York city metropolitan area using hierarchical bayesian model estimates[J]. Environment research,2016,151:399-409.
[27]CHIOU J H,PING H K. A Deep CNN-LSTM Model for Particulate Matter(PM2.5)Forecasting in Smart Cities[J]. Sensors,2018,18(7):2220-2241.
[28]CONG W,SHU L,XIAO J Y,et al. A novel spatiotemporal convolutional long short-term neural network for air pollution prediction[J]. Science of the total environment,2019,654:1091-1099.
[29]PING W S,JIA W C,JEN W H. Adaptive deep learning-based air quality prediction model using the most relevant spatial-temporal relations[J]. IEEE access 6,neurocomputing,2018(6):38186-38200.
[30]CHU D A,KAUFMAN Y J,ZIBORDI G,et al. Global monitoring of air pollution over land from the Earth observing system-terra moderate resolution imaging spectroradiometer(MODIS)[J]. Geophysis research,2003,108(D21):4661-4667.
[31]KOELEMEIJER R,HOMAN C,MATTHIJSEN J. Comparison of spatial and temporal variations of aerosol optical thickness and particulate matter over Europe[J]. Atmosphere environment,2013,40(27):5304-5315.
[32]SAIDE P E,CARMICHAEL G R,SPAK S N,et al. Forecasting urban PM10 and PM2.5 pollution episodes in very stable nocturnal conditions and complex terrain using WRF-Chem CO tracer model[J]. Atmosphere environment,2011,45(16):2769-2780.
[33]SU X,GOUGH W,SHEN Q. Correlation of pm 2.5 and meteorological variables in Ontario cities:statistical downscaling method coupled with artificial neural network[C]//Proceedings of the 24th International Conference on Modeling,Monitoring and Management of Air Pollution(AIR 2016). Crete,2016.
[34]LIANG Z,GUANG Z,PEI Y S. Learning Spatiotemporal Features using 3DCNN and Convolutional LSTM for Gesture Recognition[C]//Proceedings of the International Conference on Computer Vision. Venice,2017.
[35]DU T,BOURDEV L,FERGUS R,et al. Learning spatiotemporal features with 3D convolutional networks[C]//Proceedings of the International Conference on Computer Vision. Santiago,2015.
[36]STUART G,ELIE B,RENE D. Neural networks and the bias/variance dilemma[J]. Neural computer,1992,41(1):1-58.
[37]WIKRAM R,PAVAN Y,SHRESTHA M. Deep air:forecasting air pollution in Beijing,China.[C]//Proceedings of IEEE International Conference on Computer Vision. New York:IEEE Press,2017:1-9.
[38]赵文芳,王京丽,尚敏,等. 基于粒子群优化和支持向量机的花粉浓度预测模型[J]. 计算机应用,2019,39(1):98-104.

相似文献/References:

[1]梁星星,黄魁华,马 扬,等.周界防护的最优替换调度[J].南京师范大学学报(自然科学版),2019,42(03):52.[doi:10.3969/j.issn.1001-4616.2019.03.007]
　Liang Xingxing,Huang Kuihua,Ma Yang,et al.Optimal Replacement Scheduling for Perimeter Guarding[J].Journal of Nanjing Normal University(Natural Science Edition),2019,42(03):52.[doi:10.3969/j.issn.1001-4616.2019.03.007]
[2]闫 贺,朱 丽,戚 湧.基于改进SVM的城市道路短时交通状态预测[J].南京师范大学学报(自然科学版),2019,42(03):129.[doi:10.3969/j.issn.1001-4616.2019.03.017]
　Yan He,Zhu Li,Qi Yong.Short-term Traffic Condition Prediction ofUrban Road Based on Improved SVM[J].Journal of Nanjing Normal University(Natural Science Edition),2019,42(03):129.[doi:10.3969/j.issn.1001-4616.2019.03.017]
[3]张旭辉,张 郴,李雅南,等.城市旅游餐饮体验的注意力机制模型建构——基于机器学习的网络文本深度挖掘[J].南京师范大学学报(自然科学版),2022,45(01):32.[doi:10.3969/j.issn.1001-4616.2022.01.006]
　Zhang Xuhui,Zhang Chen,Li Yanan,et al.Construction of Attention Mechanism Model of Urban Tourism Catering Experience:Deep Mining of Online Text Based on Machine Learning[J].Journal of Nanjing Normal University(Natural Science Edition),2022,45(03):32.[doi:10.3969/j.issn.1001-4616.2022.01.006]
[4]项晓宇,朱敏捷,周灵刚,等.基于机器学习的短期规上行业工业增加值预测[J].南京师范大学学报(自然科学版),2023,46(02):99.[doi:10.3969/j.issn.1001-4616.2023.02.013]
　Xiang Xiaoyu,Zhu Minjie,Zhou Linggang,et al.Short-Term Industrial Added Value Prediction of the Above-Scale Industry Based on Machine Learning[J].Journal of Nanjing Normal University(Natural Science Edition),2023,46(03):99.[doi:10.3969/j.issn.1001-4616.2023.02.013]
[5]赵宇奔,王鑫宁,李 崇.基于K-XGBoost融合模型的高校学生学情预测研究[J].南京师范大学学报(自然科学版),2023,46(03):89.[doi:10.3969/j.issn.1001-4616.2023.03.012]
　Zhao Yuben,Wang Xingning,Li Chong.Research on Undergraduate Academic Prediction Based on K-XGBoost Fusion Model[J].Journal of Nanjing Normal University(Natural Science Edition),2023,46(03):89.[doi:10.3969/j.issn.1001-4616.2023.03.012]

备注/Memo

备注/Memo:
收稿日期:2019-07-05.基金项目:国家自然科学基金(41575156)、中国气象局软科学研究重点课题(2019ZDIANXM19). 通讯联系人:唐伟,博士,高级工程师,研究方向:气象数据分析、模式识别、深度学习. E-mail:weitang@cma.gov.cn
更新日期/Last Update: 2019-09-30