高级检索

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于渐进DS证据理论的空气质量评价方法

孙俏 张彤 王新阳 陈志泊

孙俏, 张彤, 王新阳, 陈志泊. 基于渐进DS证据理论的空气质量评价方法[J]. 北京林业大学学报. doi: 10.12171/j.1000-1522.20210228
引用本文: 孙俏, 张彤, 王新阳, 陈志泊. 基于渐进DS证据理论的空气质量评价方法[J]. 北京林业大学学报. doi: 10.12171/j.1000-1522.20210228
Sun Qiao, Zhang Tong, Wang Xinyang, Chen Zhibo. Comprehensive air quality evaluation method based on progressive strategy of DS evidence theory[J]. Journal of Beijing Forestry University. doi: 10.12171/j.1000-1522.20210228
Citation: Sun Qiao, Zhang Tong, Wang Xinyang, Chen Zhibo. Comprehensive air quality evaluation method based on progressive strategy of DS evidence theory[J]. Journal of Beijing Forestry University. doi: 10.12171/j.1000-1522.20210228

基于渐进DS证据理论的空气质量评价方法

doi: 10.12171/j.1000-1522.20210228
基金项目: 国家自然基金面上项目(32071775)
详细信息
    作者简介:

    孙俏,博士,副教授。主要研究方向:大数据技术。 Email:sunqiao0608@163.com 地址:100083北京市海淀区北京林业大学信息学院

    责任作者:

    陈志泊,博士,教授。主要研究方向:生态监测大数据、“互联网+”生态站。Email:zhibo@bjfu.edu.cn 地址:同上

Comprehensive air quality evaluation method based on progressive strategy of DS evidence theory

  • 摘要:   目的  科学地评价环境空气质量对大气污染防治具有重大的现实意义。环境空气质量的好坏是一个不确定性问题,常用的模糊综合评判法综合了多种大气污染物来评价空气质量,但评价结果往往低于我国AQI标准等级,弱化了多种污染物同时超标对空气质量的影响。DS证据理论在处理模糊或不确定等问题时具有很大优势,但是目前较少有DS证据理论评价空气质量的研究。DS证据理论的关键问题是基本概率分配(BPA)和证据冲突可能引起反直觉的结果。  方法  因此本文首先利用Pearson和Spearman相关系数分析了大气污染物浓度数据,确立了高可信度的非线性BPA函数,并提出渐进式DS证据理论策略(Pro-DS)避免了冲突证据的融合问题,建立了环境空气质量综合评价模型。然后将该模型实际应用于2019年天津市空气质量评价,并以国家标准AQI为评价指标,与主因素型与加权型常用的模糊综合评判模型对比。最后本文提出了综合污染相对系数(CPRC)量化了多种污染物对总体空气质量的影响。  结果  以AQI为评价指标,本文模型F1-score比常用的模糊综合评判模型最少提高了4.58%,最大提高了27.46%,验证了本文模型的优越性。由于AQI标准空气质量评价取决于某个污染物而不是综合多个污染物,几种模糊综合评判模型的F1-score都低于50%。以CPRC为评价指标,AQI评价结果的F1-score超过99.1%,几类模糊综合评判模型F1-score都超过89.0%,验证了CPRC的有效性。而Pro-DS模型F1-score为93.1%,综合评价空气质量的模型最优,具有很强的实际应用价值。  结论  Pro-DS模型的综合评价结果低于或高于AQI评价级别,更好地体现了多种污染物对空气质量的综合影响。相比AQI评价方式,本文Pro-DS模型得出了综合空气质量级别的概率值,CPRC指标能够对综合空气质量日排名,为相关部门预防治理空气污染和生态环境工程建设提供实质性参考。

     

  • 图  1  天津市2019年每日空气污染状况

    Figure  1.  Daily air pollution in Tianjin in 2019

    图  2  天津市2019年4个季度的日空气质量

    Figure  2.  Daily air quality in Tianjin in the four quarters of 2019

    图  3  分别在AQI和CPRC值排序总体污染状况下的AQI 评价等级与IAQI和变化趋势

    Figure  3.  The trend of the AQI level and the sum of IAQI when sorted by AQI value and CPRC respectively

    图  4  按CPRC排序后模型评价结果比较

    Figure  4.  Comprehensive assessment results of models sorted by CPRC

    图  5  模型precision、recall和F1-score比较

    Figure  5.  Comparison of precision,recall and F1-score among models

    表  1  大气污染物原始质量浓度数据统计特征

    Table  1.   Statistical characteristics of the original pollutant concentration data

    大气污染物
    Pollutant
    最大值
    Max. value
    均值
    Mean
    标准差
    SD
    SO2/(μg·m−3 60.000 0 11.041 8 5.958 6
    NO2/(μg·m−3 128.000 0 42.327 9 19.891 6
    CO/(mg·m−3 4.300 0 0.959 7 0.471 2
    O3/(μg·m−3 300.000 0 64.554 8 52.872 6
    PM10/(μg·m−3 555.000 0 80.750 5 50.099 5
    PM2.5/(μg·m−3 304.000 0 51.370 3 39.299 3
    下载: 导出CSV

    表  2  大气污染物Spearman相关系数

    Table  2.   Spearman correlation coefficient of air pollutants

    大气污染物
    Pollutant
    SO2NO2COO3PM10PM2.5
    SO2 1 0.58 0.51 −0.26 0.55 0.44
    NO2 0.58 1 0.55 −0.54 0.52 0.49
    CO 0.51 0.55 1 −0.22 0.57 0.71
    O3 −0.26 −0.54 −0.22 1 −0.10 −0.09
    PM10 0.55 0.52 0.57 −0.10 1 0.78
    PM2.5 0.44 0.49 0.71 −0.09 0.78 1
    下载: 导出CSV

    表  3  空气质量综合评价等级描述

    Table  3.   Comprehensive evaluation level of air quality

    等级
    Level
    类别
    Quality
    对健康影响
    Empact on health

    Good
    基本无影响
    Little impact

    Regular
    极弱影响
    Weak impact
    轻度
    Light
    较小影响
    Minor impact
    中度
    Moderate
    较大影响
    Greater impact
    重度
    Heavy
    对健康人群普遍有害
    Harmful
    严重
    Hazardous
    对健康人群极度有害
    Extremely harmful
    下载: 导出CSV

    表  4  {好,中,差}空气质量浓度限值

    Table  4.   Mass concentration limits of {good, medium, poor}

    大气污染物
    Pollutant

    Good

    Medium

    Poor
    SO2/(μg·m−3 50 125 350
    NO2/(μg·m−3 40 80 160
    CO/(mg·m−3 2 4 8
    O3/(μg·m−3 100 160 265
    PM10/(μg·m−3 50 150 350
    PM2.5/(μg·m−3 25 75 150
    下载: 导出CSV

    表  5  差等空气质量浓度限值

    Table  5.   Concentration limits of {Ⅳ,Ⅴ,Ⅵ} in poor air quality

    大气污染物
    Pollutant
    SO2 125 350 800
    NO2 80 160 280
    CO 4 8 12
    O3 160 265 400
    PM10 150 350 500
    PM2.5 75 150 350
    下载: 导出CSV
  • [1] Li R R, Dong Y Q, Zhu Z J, et al. A dynamic evaluation framework for ambient air pollution monitoring[J]. Applied Mathematical Modelling, 2019, 65: 52−71. doi: 10.1016/j.apm.2018.07.052
    [2] 郑霞, 胡东滨, 李权, 等. 基于组合赋权集对分析的空气质量评价:以长沙市为例[J]. 安全与环境工程, 2021, 28(1):226−232.

    Zheng X, Hu D B, Li Q, et al. Air quality assessment based on combined weighted set analysis in Changsha city[J]. Safe and Environmental Engineering, 2021, 28(1): 226−232.
    [3] Tan X R, Han L J, Zhang X Y, et al. A review of current air quality indexes and improvements under the multi-contaminant air pollution exposure[J/OL]. Journal of Environmental Management, 2021, 279: 111681[2021−04−11]. https://doi.org/10.1016/j.jenvman.2020.111681.
    [4] Li W J, Shao L Y, Wang W H, et al. Air quality improvement in response to intensified control strategies in Beijing during 2013–2019[J/OL]. Science of the Total Environment, 2020, 744: 140776[2021−03−18]. https://doi.org/10.1016/j.scitotenv.2020.140776.
    [5] Zhu C H, Li N P. Study on grey clustering model of indoor air quality indicators[J]. Procedia Engineering, 2017, 205: 2815−2822. doi: 10.1016/j.proeng.2017.09.895
    [6] Li R R, Jin Y. The early-warning system based on hybrid optimization algorithm and fuzzy synthetic evaluation model[J]. Information Sciences, 2018, 435: 296−319. doi: 10.1016/j.ins.2017.12.040
    [7] 吕铃钥, 李洪远. 基于模糊综合评判的京津冀地区空气质量评价[J]. 南开大学学报(自然科学版), 2016, 49(1):62−68.

    Lü L Y, Li H Y. Air quality evaluation of Beijing-Tianjin-Hebei region of China based on the fuzzy comprehensive evaluation method[J]. Acta Scientiarum Naturalium Universitatis Nankaiensis, 2016, 49(1): 62−68.
    [8] Xu Y Z, Du P, Wang J Z. Research and application of a hybrid model based on dynamic fuzzy synthetic evaluation for establishing air quality forecasting and early warning system: a case study in China[J]. Environmental Pollution, 2017, 223: 435−448. doi: 10.1016/j.envpol.2017.01.043
    [9] 王治朋, 王静飞, 李菊倩, 等. 基于区间数的空气质量模糊综合评价模型[J]. 数学的实践与认识, 2021, 51(3):230−235.

    Wang Z P, Wang J F, Li J Q, et al. A fuzzy comprehensive evaluation method for air quality based on interval number[J]. Mathematics in Practice and Theory, 2021, 51(3): 230−235.
    [10] 王光辉, 刘秋梅. 基于模糊综合评判的曹妃甸区环境空气质量评价[J]. 环境与可持续发展, 2018, 43(5):42−44. doi: 10.3969/j.issn.1673-288X.2018.05.011

    Wang G H, Liu Q M. Air quality evaluation of Caofeidian district by fuzzy comprehensive evaluation method[J]. Environment and Sustainable Development, 2018, 43(5): 42−44. doi: 10.3969/j.issn.1673-288X.2018.05.011
    [11] 夏冰莹. 基于决策粗糙集与证据理论的空气质量评价研究[D]. 镇江: 江苏科技大学, 2020.

    Xia B Y. Research on air quality evaluation based on decision rough set and evidence theory[D]. Zhenjiang: Jiangsu University of Science and Technology, 2020.
    [12] Ma W J, Jiang Y C, Luo X D. A flexible rule for evidential combination in Dempster–Shafer theory of evidence[J/OL]. Applied Soft Computing Journal, 2019, 85: 105512[2021−04−15]. https://doi.org/10.1016/j.asoc.2019.105512.
    [13] 田明明, 叶继华, 万叶晶. 一种迭代修正的冲突证据改进方法[J]. 控制工程, 2021, 28(3):565−570.

    Tian M M, Ye J H, Wan Y J. An improved method of iterative correction to conflict evidence[J]. Control Engineering of China, 2021, 28(3): 565−570.
    [14] 陈哲, 王加阳. 基于冲突关系网络的证据合成方法研究[J]. 电子学报, 2021, 49(1):125−131. doi: 10.12263/DZXB.20190086

    Chen Z, Wang J Y. The research on evidence combination method based on conflict relation network[J]. Acta Electronica Sinica, 2021, 49(1): 125−131. doi: 10.12263/DZXB.20190086
    [15] Yu H, Chen L Y, Yao J T. A three-way density peak clustering method based on evidence theory[J/OL]. Knowledge-Based Systems, 2021, 211: 106532 [2021-04-18]. https://doi.org/10.1016/j.knosys.2020.106532.
    [16] Chatterjee M, Namin A S. A fuzzy Dempster–Shafer classifier for detecting Web spams[J/OL]. Journal of Information Security and Applications, 2021, 59: 102793[2021−04−19]. https://doi.org/10.1016/j.jisa.2021.102793.
    [17] 肖会敏, 刘云鸽, 翟婷, 等. 基于模糊-证据理论的突发事件应急预案评估[J]. 数学的实践与认识, 2021, 51(2): 321-328

    Xiao H M, Liu Y H, Zhai T, et al. Evaluation of emergency plans based on fuzzy-evidence theory[J]. 2021, 51(2): 321-328.
    [18] Zhou Y, Li S J. BP neural network modeling with sensitivity analysis on monotonicity based Spearman coefficient[J/OL]. Chemometrics and Intelligent Laboratory Systems, 2020, 200: 103977 [2021−05−18]. https://doi.org/10.1016/j.chemolab.2020.103977.
    [19] Bappy M M, Ali S M, Kabir G, et al. Supply chain sustainability assessment with Dempster-Shafer evidence theory: Implications in cleaner production Journal of Cleaner Production[J/OL]. Journal of Cleaner Production, 2019, 237: 117771[2021−04−28]. https://doi.org/10.1016/j.jclepro.2019.117771.
    [20] Zhang S X, Lin J H, Su Let al. pDHS-DSET: Prediction of DNase I hypersensitive sites in plant genome using DS evidence theory[J]. Analytical Biochemistry, 2019, 564-565: 54−63. doi: 10.1016/j.ab.2018.10.018
    [21] Zheng T, Xu P, Denœux T. An evidential classifier based on Dempster-Shafer theory and deep learning[J]. Neurocomputing, 2021, 450: 275−293. doi: 10.1016/j.neucom.2021.03.066
  • 加载中
图(5) / 表(5)
计量
  • 文章访问数:  48
  • HTML全文浏览量:  12
  • PDF下载量:  10
  • 被引次数: 0
出版历程
  • 收稿日期:  2021-06-15
  • 修回日期:  2021-06-28
  • 网络出版日期:  2021-09-22

目录

    /

    返回文章
    返回