基于长短期记忆网络-卷积神经网络(LSTM-CNN)的北京市PM<sub>2.5</sub>浓度预测

于伸庭; 刘萍

doi:10.13205/j.hjgc.202006029

基于长短期记忆网络-卷积神经网络(LSTM-CNN)的北京市PM_2.5浓度预测

doi: 10.13205/j.hjgc.202006029

于伸庭¹,
刘萍^1,2, ,

1. 上海交通大学中英国际低碳学院, 上海 200240;
2. 上海交通大学环境科学与工程学院, 上海 200240

基金项目:

国家自然科学基金面上项目(41975152)。

详细信息

作者简介:
于伸庭(1995-),男,硕士,主要研究方向为大数据分析与数据挖掘。sjtu_yust@163.com

通讯作者:
刘萍(1976-),女,博士,主要研究方向为气溶胶形成机制及空气质量的模拟与预测。ping_liu@sjtu.edu.cn

计量
- 文章访问数: 412
- HTML全文浏览量: 51
- PDF下载量: 3
- 被引次数: 0
出版历程
- 收稿日期: 2019-09-25

LONG SHORT-TERM MEMORY-CONVOLUTION NEURAL NETWORK (LSTM-CNN) FOR PREDICTION OF PM_2.5 CONCENTRATION IN BEIJING

YU Shen-ting¹,
LIU Ping^{1,2
, ,}

1. China-UK Low Carbon College, Shanghai Jiao Tong University, Shanghai 200240, China;
2. School of Environmental Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, China

摘要

摘要: 准确预测PM_2.5浓度可以有效避免重污染天气对人体带来的危害。现有方法往往重视本地历史信息对PM_2.5浓度预测的影响,而忽略空间传输的作用。提出了一种长短期记忆网络和卷积神经网络(LSTM-CNN)相结合的方法,利用历史PM_2.5浓度数据、历史气象数据和时间数据,对空气质量监测站未来6 h PM_2.5浓度做出预测。该模型主要由2部分组成:1)基于长短期记忆网络的时序预测模型,模拟本地因素对PM_2.5浓度预测的影响;2)基于一维卷积神经网络的特征提取模型,模拟周边地区污染物的传输与扩散对PM_2.5浓度预测的影响。随机选取了北京市市区及郊区7个监测站在2014-05-01—2015-04-30期间的数据,用于研究和评估LSTM-CNN模型。结果表明:提出的LSTM-CNN模型相比于LSTM模型具有更好的预测效果,且对于郊区站点预测效果的改进略优于市区站点。
- LSTM网络 /
- CNN网络 /
- PM_2.5浓度预测 /
- 深度学习 /
- 时空数据
Abstract: The prediction of PM_2.5 can effectively prevent people from the harm by heavy pollution. However, the existing methods often emphasize the influence of local historical information and neglect the effect of spatial transport. In this paper, we proposed a method, called as long-short-term memory-convolutional neural network (LSTM-CNN), to predict PM_2.5 concentration of a specific air quality monitoring station over 6 h using historical PM_2.5 concentration data, historical weather data, and time stamp data. The model consisted of two parts: 1) using long-short-term memory networks to model the local variation of PM_2.5 concentrations caused by local factors; 2) using one-dimensional convolutional neural networks to model the variation of PM_2.5 concentrations caused by spatial transport. We randomly selected 7 monitoring stations in urban and rural areas in Beijing from May 1^st 2014 to April 30^th 2015 to conduct the evaluation of LSTM-CNN model. The results showed that the proposed LSTM-CNN model could provide a better prediction result than LSTM model, and a better result for monitoring stations in rural areas than those in urban areas.
- LSTM /
- CNN /
- prediction of PM_2.5 concentration /
- deep learning /
- spatiotemporal data

HTML全文

参考文献(22)

KIOUMOURTZOGLOU M A,SCHWARTZ J,JAMES P,et al. PM_2.5 and mortality in 207 US cities modification by temperature and city characteristics[J]. Epidemiology,2016,27(2): 221-227.

中华人民共和国环保部.2015中国环境状况公报[Z]. 北京:[2016-06-01

CHEN J 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.

WANG Z,MAEDA T,HAYASHI M,et al. A nested air quality prediction modeling system for urban and regional scales: application for high-ozone episode in Taiwan[J]. Water Air & Soil Pollution,2001,130(1/2/3/4): 391-396.

SAIDE P E,CARMICHAEL G R,SPAK S N,et al. Forecasting urban PM₁₀ and PM_2.5 pollution episodes in very stable nocturnal conditions and complex terrain using WRF-Chem CO tracer model[J]. Atmospheric Environment,2011,45(16): 2769-2780.

LI X,PENG L,YAO X J,et al. Long short-term memory neural network for air pollutant concentration predictions: method development and evaluation[J]. Environmental Pollution,2017,231(1): 997-1004.

陈宁,毛善君,李德龙,等. 多基站协同训练神经网络的PM_2.5预测模型[J]. 测绘科学,2018,241(7): 87-93.

BOX G E,JENKINS G M. Time series analysis: forecasting and control rev. ed[J]. Journal of Time,1976,31(4): 238-242.

侯俊雄,李琦,朱亚杰,等. 基于随机森林的PM_2.5实时预报系统[J]. 测绘科学,2017,42(1): 1-6.

GARCÍA NIETO P J,COMBARRO E F,DEL COZ DÍAZ J J,et al. A SVM-based regression model to study the air quality at local scale in Oviedo urban area (Northern Spain): a case study[J]. Applied Mathematics & Computation,2013,219(17): 8923-8937.

HOOYBERGHS J,MENSINK C,DUMONT G,et al. A neural network forecast for daily average PM₁₀ concentrations in Belgium[J]. Atmospheric Environment,2005,39(18): 3279-3289.

YU F,ZHANG W F,SUN D Z,et al. Ozone concentration forecast method based on genetic algorithm optimized back propagation neural networks and support vector machine data classification[J]. Atmospheric Environment,2011,45(11): 1979-1985.

ZHAO J C,DENG F,CAI Y Y,et al. Long short-term memory-Fully connected (LSTM-FC) neural network for PM_2.5 concentration prediction[J]. Chemosphere,2019,220: 486-492.

HUANG C J,KUO P H. A deep CNN-LSTM model for particulate matter (PM_2.5) forecasting in smart cities[J]. Sensors,2018,18(7): 2220-0000.

LIU X D,LIU Q,ZOU Y Y,et al. A self-organizing LSTM-based approach to PM_2.5 forecast[C]//International Conference on Cloud Computing and Security.ACM,2018.

ZHENG Y,LIU F R,HSIEH H P. U-Air: when urban air quality inference meets big data[C]//Acm Sigkdd International Conference on Knowledge Discovery & Data Mining.ACM,2013.

ZHENG Y,CAPRA L,WOLFSON O,et al. Urban computing: concepts, methodologies, and applications[J]. Acm Transactions on Intelligent Systems & Technology,2014,5(3): 1-2.

ZHENG Y,YI X W,LI M,et al. Forecasting fine-grained air quality based on big data[C]//Acm Sigkdd International Conference on Knowledge Discovery & Data Mining.ACM,2015.

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]//Longhurst J W S, Brebbia C A,Barnes J.24th International Conference on Modelling, Monitoring and Management of Air Pollution,Greece,2016:215-226.

张佳华,侯英雨,李贵才,等. 北京城市及周边热岛日变化及季节特征的卫星遥感研究与影响因子分析[J]. 中国科学:地球科学,2005,35(增刊1): 187-194.

ZHANG Y X,ZHANG Y M,WANG Y S,et al. PIXE characterization of PM₁₀ and PM_2.5 particulate matter collected during the winter season in Shanghai city[J].Journal of Radioanalytical & Nuclear Chemistry,2006,267(2): 497-499.

王清川,周贺玲,许敏,等. 河北省廊坊市大气污染扩散气象条件影响分析[J]. 防灾科技学院学报,2014,16(3): 1-8.

施引文献

资源附件(0)

访问统计

点击查看大图

计量

文章访问数: 412
HTML全文浏览量: 51
PDF下载量: 3
被引次数: 0

姓名
邮箱
手机号码
标题
留言内容
验证码

留言板

基于长短期记忆网络-卷积神经网络(LSTM-CNN)的北京市PM_2.5浓度预测

doi: 10.13205/j.hjgc.202006029

作者简介:
于伸庭(1995-),男,硕士,主要研究方向为大数据分析与数据挖掘。sjtu_yust@163.com

通讯作者:
刘萍(1976-),女,博士,主要研究方向为气溶胶形成机制及空气质量的模拟与预测。ping_liu@sjtu.edu.cn

计量

LONG SHORT-TERM MEMORY-CONVOLUTION NEURAL NETWORK (LSTM-CNN) FOR PREDICTION OF PM_2.5 CONCENTRATION IN BEIJING

计量

目录

期刊在线

作者中心

友情链接

留言板

基于长短期记忆网络-卷积神经网络(LSTM-CNN)的北京市PM2.5浓度预测

doi: 10.13205/j.hjgc.202006029

作者简介: 于伸庭(1995-),男,硕士,主要研究方向为大数据分析与数据挖掘。sjtu_yust@163.com

通讯作者: 刘萍(1976-),女,博士,主要研究方向为气溶胶形成机制及空气质量的模拟与预测。ping_liu@sjtu.edu.cn

计量

出版历程

LONG SHORT-TERM MEMORY-CONVOLUTION NEURAL NETWORK (LSTM-CNN) FOR PREDICTION OF PM2.5 CONCENTRATION IN BEIJING

计量

出版历程

目录

期刊在线

作者中心

友情链接

基于长短期记忆网络-卷积神经网络(LSTM-CNN)的北京市PM_2.5浓度预测

作者简介:
于伸庭(1995-),男,硕士,主要研究方向为大数据分析与数据挖掘。sjtu_yust@163.com

通讯作者:
刘萍(1976-),女,博士,主要研究方向为气溶胶形成机制及空气质量的模拟与预测。ping_liu@sjtu.edu.cn

LONG SHORT-TERM MEMORY-CONVOLUTION NEURAL NETWORK (LSTM-CNN) FOR PREDICTION OF PM_2.5 CONCENTRATION IN BEIJING