南京大学学报(自然科学版) ›› 2020, Vol. 56 ›› Issue (1): 142–150.doi: 10.13232/j.cnki.jnju.2020.01.016

• • 上一篇    下一篇

基于社区划分的空气质量指数(AQI)预测算法

袁燕,陈伯伦(),朱国畅,花勇,于永涛   

  1. 淮阴工学院计算机与软件工程学院,淮安,223003
  • 收稿日期:2019-08-13 出版日期:2020-01-30 发布日期:2020-01-10
  • 通讯作者: 陈伯伦 E-mail:chenbolun1986@163.com
  • 基金资助:
    国家自然科学基金(61602202);国家重点研发项目(2018YFB1004904);江苏省自然科学基金(BK20160428);江苏省六大人才高峰项目(XYDXX?034)

Prediction of Air Quality Index (AQI) based on community division

Yan Yuan,Bolun Chen(),Guochang Zhu,Yong Hua,Yongtao Yu   

  1. College of Computer and Software Enginnering,Huaiyin Institute of Technology,Huaian,223003,China
  • Received:2019-08-13 Online:2020-01-30 Published:2020-01-10
  • Contact: Bolun Chen E-mail:chenbolun1986@163.com

摘要:

AQI (Air Quality Index)是判定空气质量好坏的重要指标,做好AQI的预测,对大气污染的治理有积极的推进作用,但目前预测AQI的算法通常没有综合考虑气象因素和周边城市对预测性能的影响.将气象因素和周边城市的污染物因素作为算法设计的基础,提出一种基于社区划分的空气质量指数预测的算法.首先根据气象特征计算城市之间的相似度,接着对各城市间的相似度矩阵进行社区划分;然后将属于同一社区的城市污染物时序信息作为预测目标城市空气质量指数的依据,并考虑目标城市的周边城市对其的影响;最后使用非线性回归的方法进行预测建模.通过对江苏省内20座城市的大气污染数据和气象数据的采集与分析,证明该算法不但预测精度有所提高,而且与传统的时间序列预测模型相比,降低了时间复杂度.

关键词: 空气质量指数(AQI)预测, 气象因素, 时序信息, 社区划分

Abstract:

AQI (Air Quality Index) is an important indicator to judge the air quality. Effectively predicting the AQI has positive impact on the control of air pollution. However,the existing AQI prediction methods scarcely consider the weather factors and the influence on the prediction performance of the surrounding cities. In this paper,we propose a community division based AQI prediction method by considering the weather factors and the pollutant factors of the surrounding cities. Firstly,the similarity between cities is computed according to the weather factors. Then,community division is performed on the similarity matrices of each pair of cities. Next,by considering the impact of the surrounding cities of the target city,the city pollutant time series information belonging to the same community is treated as the basis for predicting the AQI of the target city. Finally,nonlinear regression is conducted for predictive modelling. Through the collection and analysis on the air pollution data and weather data of 20 cities in Jiangsu Province,it demonstrates that the proposed method improves greatly in prediction accuracy and performs computational effectively compared with the traditional time series based prediction models.

Key words: Air Quality Index (AQI), weather factors, time series information, community division

中图分类号: 

  • TP301.6

表1

空气质量分指数及对应的污染物项目浓度限值"

IAQIContaminant project concentration limit
SO2 (μg·m-1)NO2 (μg·m-1)PM10 (μg·m-1)CO (μg·m-1)O3 (μg·m-1)PM2.5 (μg·m-1)
0000000
50504050210035
10015080150416075
15047518025015215115
20080028035024265150
300160056542036800250
400210075050048350
500262094060060500

表2

符号介绍"

符号描述
IAQIA污染物因素A的空气质量分指数
CA污染物因素A的质量浓度值
BPHi表1中与CA相近的污染物浓度限值的高位值
BPLo表1中与CA相近的污染物浓度限值的低位值
IAQIHi表1中与BPHi对应的空气质量分指数
IAQILo表1中与BPLo对应的空气质量分指数
ui样本中编号为i的城市
dt样本的天数t
xtt天的气象因素X的值
ytt天的污染因素Y的值
Bit(X)江苏省第i个城市第t天第X个气象因素的值
Ait(Y)江苏省第i个城市第t天第Y个污染物因素的值
AQIit预测城市it天AQI的值
sim(Bm',Bn')任意两个城市mn间气象因素的相似度矩阵
k表示第k种气象因素
Bi'表示气象因素数据矩阵
Ai'表示污染物因素数据矩阵
Ah(Y)表示目标城市h关于污染物因素Y的矩阵
SA(m,n)城市mn间的相似度矩阵

图1

关于气象因素的数学模型"

图2

关于污染因素的数学模型"

图3

社区划分的示例"

图4

CO,NO2,O3,PM10,PM2.5和SO2在未来六天的预测值"

图5

CO,NO2,O3,PM10,PM2.5和SO2的预测值在六天内跟实际值的误差"

表3

第七天AQI的预测值"

AlgorithmCONO2O3PM10PM2.5SO2AQI
Actual value141581377610102
CK?UNR163.2662.48105.387812.75103.79
UNR1.1141.2163.42115.76110.911.13144.89
BP0.905147127601088.5
ELM1.2160.5163.76140.1681.3914.13107.99

表4

用来评估预测模型的参数值"

AlgorithmMSERMSEMAPEMAE
CK?UNR3.201.790.0171.79
UNR1839.5542.890.4242.89
BP182.2513.50.1313.5
ELM35.885.990.0585.99
1 Thach T Q,Tsang H,Cao P H,et al. A novel method to construct an air quality index based on air pollution profiles. International Journal of Hygiene and Environmental Health,2017,221(1):17-26.
2 Brunekreef B,Holgate S T. Air pollution and health. The Lancet,2002,360(9341):1233-1242.
3 张欣,许建明,王体健等. 上海市一次重霾污染过程的特征及成因分析. 南京大学学报(自然科学),2015(3):463-472.
Zhang X,Xu J M,Wang T J,et al. Characteristics and formation mechanism of a serious haze episode in December 2013 in Shanghai. Journal of Nanjing University (Natural Science),2015(3):463-472.
4 Fu B,Xiao H G,Wu L F. Grey relational analysis for the AQI of Beijing,Tianjin,and Shijiazhuang and related countermeasures. Grey Systems:Theory and Application,2018,8(2):156-166.
5 Yao W X,Zhang C X,Xiao W,et al. The research of new daily diffuse solar radiation models modified by air quality index (AQI) in the region with heavy fog and haze. Energy Conversion and Management,2017,139:140-150.
6 Liu C M. Effect of PM2.5 on AQI in Taiwan. Environmental Modelling & Software,2002,17(1):29-37.
7 Carbajal?Hernández J J,Sánchez?Fernández L P,Carrasco?Ochoa J A,et al. Assessment and prediction of air quality using fuzzy logic and autoregressive models. Atmospheric Environment,2012,60:37-50.
8 Corani G,Scanagatta M. Air pollution prediction via multi?label classification. Environmental Modelling & Software,2016,80:259-264.
9 Singh K P,Gupta S,Kumar A,et al. Linear and nonlinear modeling approaches for urban air quality prediction. Science of the Total Environment,2012,426:244-255.
10 李博群,贾政权,刘利平. 基于模糊时间序列的空气质量指数预测. 华北理工大学学报(自然科学版),2018,40(3):78-86.
Li B Q,Jia Z Q,Liu L PAir quality index forecast based on fuzzy time series models. Journal of North China University of Science and Technology (Natural Science Edition),2018,40(3):78-86.
11 高帅,胡红萍,李洋等. 基于改进的思维进化算法与BP神经网络的AQI预测. 数学的实践与认识,2018,48(19):151-157.
Gao S,Hu H P,Li Y,et al. AQI prediction based on mind evolutionary algorithm and BP neural network. Mathematics in Practice and Theory,2018,48(19):151-157.
12 刘洪通,冯百明,温向慧等. 基于Storm的AQI实时预测模型. 计算机工程与设计,2019,40(1):296-301.
Liu H T,Feng B M,Wen X H,et al. Real?time AQI prediction model based on Storm. Computer Engineering and Design,2019,40(1):296-301.
13 张春露,白艳萍. 基于TensorFlow的LSTM模型在太原空气质量AQI指数预测中的应用. 重庆理工大学学报(自然科学),2018,32(8):137-141.
Zhang C L,Bai Y P.Application of LSTM prediction model based on tensor flow in taiyuan air quality AQI index. Journal of Chongqing Institute of Technology,2018,32(8):137-141.
14 白鹤鸣,沈润平,师华定等. 基于BP神经网络的空气污染指数预测模型研究. 环境科学与技术,2013,36(3):186-189.
Bai H M,Shen R P,Shi H D,et al. Forecasting model of air pollution index based on BP neural network. Environmental Science & Technology,2013,36(3):186-189.
15 Li X,Peng L,Hu Y,et al. Deep learning architecture for air quality predictions. Environmental Science and Pollution Research,2016,23(22):22408-22417.
16 Xi X,Wei Z,Rui X G,et al. A comprehensive evaluation of air pollution prediction improvement by a machine learning method∥2015 IEEE International Conference on Service Operations,Logistics and Informatics (SOLI). Hammamet,Tunisia:IEEE,2015:176-181.
17 Han L J,Zhou W Q,Li W F,et al. Meteorological and urban landscape factors on severe air pollution in Beijing. Journal of the Air & Waste Management Association,2015,65(7):782-787.
18 Sharma M,Aggarwal S,Bose P,et al. Meteorology?based forecasting of air quality index using neural network∥IEEE International Conference on Industrial Informatics. Banff, Canada:IEEE,2003:374-378.
19 Cui H,Ma R,Gao F. Relationship between meteorological factors and diffusion of atmospheric pollutants. Chemical Engineering Transactions,2018,71:1417-1422.
20 Li Y,Chiu Y H,Lu L C. Energy and AQI performance of 31 cities in China. Energy Policy,2018,122:194-202.
21 Liu M,Gu J L,Liu L J,et al. Air quality analysis of Dalian in summer based on AQI data. Journal of Atmospheric and Environmental Optics,2016,11(2):111-117.
22 Newman M E J. Fast algorithm for detecting community structure in networks. Physical Review E,2004,69(6):066133.
23 Newman M E J. Community detection in networks:. modularity optimization and maximum likelihood are equivalent2016,arXiv:1606.02319.
[1] 李亚重,杨有龙,仇海全. 一种基于嵌入式的弱标记分类算法[J]. 南京大学学报(自然科学版), 2020, 56(4): 549-560.
[2] 刘亮,何庆. 基于改进蝗虫优化算法的特征选择方法[J]. 南京大学学报(自然科学版), 2020, 56(1): 41-50.
[3] 洪思思,曹辰捷,王 喆*,李冬冬. 基于矩阵的AdaBoost多视角学习[J]. 南京大学学报(自然科学版), 2018, 54(6): 1152-1160.
[4] 周星星,张海平,吉根林. 具有时空特性的区域移动模式挖掘算法[J]. 南京大学学报(自然科学版), 2018, 54(6): 1171-1182.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
[1] 许 林,张 巍*,梁小龙,肖 瑞,曹剑秋. 岩土介质孔隙结构参数灰色关联度分析[J]. 南京大学学报(自然科学版), 2018, 54(6): 1105 -1113 .
[2] 卢 毅,于 军,龚绪龙,王宝军,魏广庆,季峻峰. 基于DFOS的连云港第四纪地层地面沉降监测分析[J]. 南京大学学报(自然科学版), 2018, 54(6): 1114 -1123 .
[3] 王 倩,聂秀山,耿蕾蕾,尹义龙. D2D通信中基于Q学习的联合资源分配与功率控制算法[J]. 南京大学学报(自然科学版), 2018, 54(6): 1183 -1192 .
[4] 孔 颉, 孙权森, 纪则轩, 刘亚洲. 基于仿射不变离散哈希的遥感图像快速目标检测新方法[J]. 南京大学学报(自然科学版), 2019, 55(1): 49 -60 .
[5] 顾健伟, 曾 诚, 邹恩岑, 陈 扬, 沈 艺, 陆 悠, 奚雪峰. 基于双向注意力流和自注意力结合的机器阅读理解[J]. 南京大学学报(自然科学版), 2019, 55(1): 125 -132 .
[6] 安 晶, 艾 萍, 徐 森, 刘 聪, 夏建生, 刘大琨. 一种基于一维卷积神经网络的旋转机械智能故障诊断方法[J]. 南京大学学报(自然科学版), 2019, 55(1): 133 -142 .
[7] 王蔚, 胡婷婷, 冯亚琴. 基于深度学习的自然与表演语音情感识别[J]. 南京大学学报(自然科学版), 2019, 55(4): 660 -666 .
[8] 王博闻, 史江峰, 史逝远, 张伟杰, 马晓琦, 赵业思. 基于遥感数据定位老龄树群[J]. 南京大学学报(自然科学版), 2019, 55(4): 699 -707 .
[9] 马益平,严浩军,王琼京,赵亚云,张秋菊,孔春龙. 混合配体法合成氨基MIL⁃101(Cr)及其二氧化碳吸附和除湿性能[J]. 南京大学学报(自然科学版), 2019, 55(5): 840 -849 .
[10] 李勤,陆现彩,张立虎,程永贤,刘鑫. 蒙脱石层间阳离子交换的分子模拟[J]. 南京大学学报(自然科学版), 2019, 55(6): 879 -887 .