基于RNN的空气污染时空预报模型研究

被引：50

作者：

范竣翔 ^{[1
]}

李琦 ^{[1
,2
]}

朱亚杰 ^{[1
]}

侯俊雄 ^{[1
]}

冯逍 ^{[1
]}

机构：

[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

来源：

测绘科学 | 2017年 / 07期

关键词：

空气污染; 缺失值; RNN; LSTM; 深度学习;

D O I：

10.16251/j.cnki.1009-2307.2017.07.013

中图分类号：

X51 [大气污染及其防治];

学科分类号：

0706 ; 070602 ;

摘要：

针对空气污染物时间序列中包含缺失值以及现有时间序列预报模型缺乏对时序特征状态建模的问题,该文构建了基于缺失值处理算法和RNN(循环神经网络)的时空预报框架。对空气污染物时序数据设计了3种缺失值处理算法(前向递补、均值替代和权重衰减),用缺失标签和缺失时长对缺失值建模,并在此基础上搭建含有全连接层与LSTM层的深度循环神经网络(DRNN)用于时空预报。使用深度全连接神经网络(DFNN)作为DRNN的对照,用京津冀区域的空气质量和气象数据训练模型,并比较不同模型的预测精度。通过实验,比较了3种缺失值处理方法的效果,结果表明,LSTM在空气污染时空序列预测上的表现优于传统的全连接神经网络层,证实了提出的基于深度学习的时空预报框架的有效性。

引用

页码：76 / 83+120 +120

页数：9

共 8 条

[1]

Forecasting SO 2 Pollution Incidents by means of Elman Artificial Neural Networks and ARIMA Models[J] . Antonio Bernardo Sánchez,Celestino Ordó?ez,Fernando Sánchez Lasheras,Francisco Javier de Cos Juez,Javier Roca-Pardi?as,L. Jódar.Abstract and Applied Analysis . 2013

[2]

Communicating air pollution-related health risks to the public: An application of the Air Quality Health Index in Shanghai, China[J] . Renjie Chen,Xi Wang,Xia Meng,Jing Hua,Zhijun Zhou,Bingheng Chen,Haidong Kan.Environment International . 2013

[3]

Intercomparison of air quality data using principal component analysis, and forecasting of PM 10 and PM 2.5 concentrations using artificial neural networks, in Thessaloniki and Helsinki[J] . Science of the Total Environment . 2011 (7)

[4] Comparison of correlation analysis techniques for irregularly sampled time series [J].

Rehfeld, K. ;

Marwan, N. ;

Heitzig, J. ;

Kurths, J. .

NONLINEAR PROCESSES IN GEOPHYSICS, 2011, 18 (03) :389-404

[5]

A New Multipollutant, No-Threshold Air Quality Health Index Based on Short-Term Associations Observed in Daily Time-Series Analyses[J] . Stieb,David M,Burnett,Richard T,Smith-Doiron,Marc,Brion,Orly,Shin,Hwashin Hyun,Economou,Vanita.Journal of the Air & Waste Management Association . 2008 (3)

[6]

Forecasting of the daily meteorological pollution using wavelets and support vector machine[J] . Stanislaw Osowski,Konrad Garanty.Engineering Applications of Artificial Intelligence . 2006 (6)

[7] Fully coupled "online" chemistry within the WRF model [J].

Grell, GA ;

Peckham, SE ;

Schmitz, R ;

McKeen, SA ;

Frost, G ;

Skamarock, WC ;

Eder, B .

ATMOSPHERIC ENVIRONMENT, 2005, 39 (37) :6957-6975

[8] Long short-term memory [J].

Hochreiter, S ;

Schmidhuber, J .

NEURAL COMPUTATION, 1997, 9 (08) :1735-1780

← 1 →