Abstract
The subway has increasingly taken over as the primary method of short-distance travel in the development of modern urban transportation. Due to the poor air mobility in subway stations, it is crucial to monitor and provide alerts on air quality. This work provides a probabilistic prediction framework of PM2.5 concentration to solve the air quality early warning problem in subway stations. Firstly, outliers are discovered and corrected utilizing a probabilistic-based auto-encoder (PAE). Secondly, the multi-resolution elastic-gated attention mechanism is used to address error accumulation and historical information lost during the prediction process. Moreover, the decoder structure of sequence to sequence (Seq2Seq) is improved through multiple output strategy and flexible gate attention mechanism to reduce error accumulation. Finally, the Seq2Seq is equipped with Gaussian mixture models (GMM), allowing it to adapt to more complicated changes and produce a probability distribution of PM2.5 concentrations. According to tests on data from subway stations, the Pinball loss and Winkler score of the proposed model are smaller compared to other models.
摘要
地铁是城市中短途通勤的主流交通工具。然而,由于地铁站内空气流动性差,所以监测地铁空 气质量并提供高污染预警具有重大意义。为解决地铁站内空气质量预警问题,本文提出了一个概率预 测框架。首先,采用基于概率损失的自动编码器对PM2.5数据进行检测并纠**异常值。其次,采用多分 辨率弹性门控注意力机制改进序列到序列(Seq2Seq)中编码器的结构,以减弱预测过程中的误差累积和 历史信息丢失。此外,将多输出策略和弹性门控注意力机制相结合,来改进Seq2Seq 中解码器的结构, 以减少预测时的误差累积。最后,在Seq2Seq 的输出端嵌入高斯混合模型(Gaussian mixture model, GMM),使其能够适应浓度分布的变化并生成浓度的概率分布。所提出的模型在第1 步、3 步和5 步预 测中的**均MAPE分别为0.94%、1.16%和1.26%。模型对地铁站PM2.5数据的测试结果表明,与其他模 型相比,所提出的模型的Pinball loss和Winkler score更小。
References
SON YS, OH Y H, CHOI I Y, et al. Development of a magnetic hybrid filter to reduce PM10 in a subway platform [J]. Journal of Hazardous Materials, 2019, 368: 197–203. DOI: https://doi.org/10.1016/j.jhazmat.2019.01.045.
XIAO De, LI Bao-xi, CHENG Shi-xiong. The effect of subway development on air pollution: Evidence from China [J]. Journal of Cleaner Production, 2020, 275: 124149. DOI: https://doi.org/10.1016/j.jclepro.2020.124149.
ZHENG Xun, LI Hai-ying, MENG Ling-yun, et al. Improved social force model based on exit selection for microscopic pedestrian simulation in subway station [J]. Journal of Central South University, 2015, 22(11): 4490–4497. DOI: https://doi.org/10.1007/s11771-015-2997-5.
SMITH J D, BARRATT B M, FULLER G W, et al. PM2.5 on the London underground [J]. Environment International, 2020, 134: 105188. DOI: https://doi.org/10.1016/j.envint.2019.105188.
PAN Song, DU Sai-sai, WANG **n-ru, et al. Analysis and interpretation of the particulate matter (PM10 and PM2.5) concentrations at the subway stations in Bei**g, China [J]. Sustainable Cities and Society, 2019, 45: 366–377. DOI: https://doi.org/10.1016/j.scs.2018.11.020.
PARCHAMI M, MOTLAGH M S, PARDAKHTI A. Particulate matter concentrations and characterization in urban subway system-case study Tehran, Iran [J]. Journal of Environmental Health Science and Engineering, 2022, 20(2): 953–962. DOI: https://doi.org/10.1007/s40201-022-00838-7.
QIAO Ting, XIU Guang-li, ZHENG Yi, et al. Preliminary investigation of PM1, PM2.5, PM10 and its metal elemental composition in tunnels at a subway station in Shanghai, China [J]. Transportation Research Part D: Transport and Environment, 2015, 41: 136–146. DOI: https://doi.org/10.1016/j.trd.2015.09.013.
LI Zhi-yuan, CHE Wen-wei, FREY H C, et al. Factors affecting variability in PM2.5 exposure concentrations in a metro system [J]. Environmental Research, 2018, 160: 20–26. DOI: https://doi.org/10.1016/j.envres.2017.09.006.
FIGUEROA-LARA J J, MURCIA-GONZÁLEZ J M, GARCÍA-MARTÍNEZ R, et al. Effect of platform subway depth on the presence of Airborne PM2.5, metals, and toxic organic species [J]. Journal of Hazardous Materials, 2019, 377: 427–436. DOI: https://doi.org/10.1016/j.jhazmat.2019.05.091.
MUGICA-ÁLVAREZ V, FIGUEROA-LARA J, ROMERO-ROMO M, et al. Concentrations and properties of airborne particles in the Mexico City subway system [J]. Atmospheric Environment, 2012, 49: 284–293. DOI: https://doi.org/10.1016/j.atmosenv.2011.11.038.
GUSEVA CANU I, CRÉZÉ C, HEMMENDINGER M, et al. Particle and metal exposure in Parisian subway: Relationship between exposure biomarkers in air, exhaled breath condensate, and urine [J]. International Journal of Hygiene and Environmental Health, 2021, 237: 113837. DOI: https://doi.org/10.1016/j.ijheh.2021.113837.
GUO Li, HU Yun-jie, HU Qing-qing, et al. Characteristics and chemical compositions of particulate matter collected at the selected metro stations of Shanghai, China [J]. Science of the Total Environment, 2014, 496: 443–452. DOI: https://doi.org/10.1016/j.scitotenv.2014.07.055.
ZHANG Meng-yuan, SHAO Long-yi, JONES T, et al. Atmospheric iron particles in PM2.5 from a subway station, Bei**g, China [J]. Atmospheric Environment, 2022, 283: 119175. DOI: https://doi.org/10.1016/j.atmosenv.2022.119175.
JI Wen-**g, LI **ao-feng, WANG Chun-wang. Composition and exposure characteristics of PM2.5 on subway platforms and estimates of exposure reduction by protective masks [J]. Environmental Research, 2021, 197: 111042. DOI: https://doi.org/10.1016/j.envres.2021.111042.
ROY D, LYOU E S, KIM J, et al. Commuters health risk associated with particulate matter exposures in subway system-Globally [J]. Building and Environment, 2022, 216: 109036. DOI: https://doi.org/10.1016/j.buildenv.2022.109036.
KIM M, SANKARARAO B, KANG O, et al. Monitoring and prediction of indoor air quality (IAQ) in subway or metro systems using season dependent models [J]. Energy and Buildings, 2012, 46: 48–55. DOI: https://doi.org/10.1016/j.enbuild.2011.10.047.
LIM J, KIM Y, OH T, et al. Analysis and prediction of indoor air pollutants in a subway station using a new key variable selection method [J]. Korean Journal of Chemical Engineering, 2012, 29(8): 994–1003. DOI: https://doi.org/10.1007/s11814-011-0278-z.
LIU Hong-bin, YANG Chong, HUANG Ming-zhi, et al. Modeling of subway indoor air quality using Gaussian process regression [J]. Journal of Hazardous Materials, 2018, 359: 266–273. DOI: https://doi.org/10.1016/j.jhazmat.2018.07.034.
FU Ming-lei, WANG Wei-wen, LE Zi-chun, et al. Prediction of particular matter concentrations by developed feed-forward neural network with rolling mechanism and gray model [J]. Neural Computing and Applications, 2015, 26(8): 1789–1797. DOI: https://doi.org/10.1007/s00521-015-1853-8.
BALRAM D, LIAN K Y, SEBASTIAN N. Air quality warning system based on a localized PM2.5 soft sensor using a novel approach of Bayesian regularized neural network via forward feature selection [J]. Ecotoxicology and Environmental Safety, 2019, 182: 109386. DOI: https://doi.org/10.1016/j.ecoenv.2019.109386.
GARCÍA NIETO P J, SÁNCHEZ LASHERAS F, GARCÍA-GONZALO E, et al. PM10 concentration forecasting in the metropolitan area of Oviedo (Northern Spain) using models based on SVM, MLP, VARMA and ARIMA: A case study [J]. Science of the Total Environment, 2018, 621: 753–761. DOI: https://doi.org/10.1016/j.scitotenv.2017.11.291.
DONG Ming, YANG Dong, KUANG Yan, et al. PM2.5 concentration prediction using hidden semi-Markov model-based times series data mining [J]. Expert Systems With Applications, 2009, 36(5): 9046–9055. DOI: https://doi.org/10.1016/j.eswa.2008.12.017.
SUN Wei, ZHANG Hao, PALAZOGLU A, et al. Prediction of 24-hour-average PM2.5 concentrations using a hidden Markov model with different emission distributions in Northern California [J]. Science of the Total Environment, 2013, 443: 93–103. DOI: https://doi.org/10.1016/j.scitotenv.2012.10.070.
WANG Jian-zhou, LI Hong-min, LU Hai-yan. Application of a novel early warning system based on fuzzy time series in urban air quality forecasting in China [J]. Applied Soft Computing, 2018, 71: 783–799. DOI: https://doi.org/10.1016/j.asoc.2018.07.030.
KARIMIAN H, LI Qi, WU Chun-lin, et al. Evaluation of different machine learning approaches to forecasting PM2.5 mass concentrations [J]. Aerosol and Air Quality Research, 2019, 19(6): 1400–1410. DOI: https://doi.org/10.4209/aaqr.2018.12.0450.
KIM B Y, LIM Y K, CHA J W. Short-term prediction of particulate matter (PM10 and PM2.5) in Seoul, South Korea using tree-based machine learning algorithms [J]. Atmospheric Pollution Research, 2022, 13(10): 101547. DOI: https://doi.org/10.1016/j.apr.2022.101547.
LIU Hui, DENG Da-hua. An enhanced hybrid ensemble deep learning approach for forecasting daily PM2.5 [J]. Journal of Central South University, 2022, 29(6): 2074–2083. DOI: https://doi.org/10.1007/s11771-022-5051-4.
LOY-BENITEZ J, TARIQ S, NGUYEN H T, et al. Neural circuit policies-based temporal flexible soft-sensor modeling of subway PM2.5 with applications on indoor air quality management [J]. Building and Environment, 2022, 207: 108537. DOI: https://doi.org/10.1016/j.buildenv.2021.108537.
PAK U, MA Jun, RYU U, et al. Deep learning-based PM2.5 prediction considering the spatiotemporal correlations: A case study of Bei**g, China [J]. Science of the Total Environment, 2020, 699: 133561. DOI: https://doi.org/10.1016/j.scitotenv.2019.07.367.
ZHANG Bo, ZHANG Han-wen, ZHAO Geng-ming, et al. Constructing a PM2.5 concentration prediction model by combining auto-encoder with Bi-LSTM neural networks [J]. Environmental Modelling & Software, 2020, 124: 104600. DOI: https://doi.org/10.1016/j.envsoft.2019.104600.
SOH P W, CHANG Jia-wei, HUANG J W. Adaptive deep learning-based air quality prediction model using the most relevant spatial-temporal relations [J]. IEEE Access, 2018, 6: 38186–38199. DOI: https://doi.org/10.1109/ACCESS.2018.2849820.
YEO I, CHOI Y, LOPS Y, et al. Efficient PM2.5 forecasting using geographical correlation based on integrated deep learning algorithms [J]. Neural Computing and Applications, 2021, 33(22): 15073–15089. DOI: https://doi.org/10.1007/s00521-021-06082-8.
ZHANG Luo, LIU Peng, ZHAO Lei, et al. Air quality predictions with a semi-supervised bidirectional LSTM neural network [J]. Atmospheric Pollution Research, 2021, 12(1): 328–339. DOI: https://doi.org/10.1016/j.apr.2020.09.003.
YANG Bai-lin, SUN Shu-lin, LI Jian-yuan, et al. Traffic flow prediction using LSTM with feature enhancement [J]. Neurocomputing, 2019, 332: 320–327. DOI: https://doi.org/10.1016/j.neucom.2018.12.016.
BAHDANAU D, CHO K, BENGIO Y. Neural machine translation by jointly learning to align and translate [EB/OL]. 2014. ar**v: 1409.0473. https://arxiv.org/abs/1409.0473.
BI **g, LIN Yong-ze, DONG Quan-xi, et al. An improved attention-based LSTM for multi-step dissolved oxygen prediction in water environment [C]// 2020 IEEE International Conference on Networking, Sensing and Control (ICNSC). IEEE, 2020: 1–6. DOI: https://doi.org/10.1109/ICNSC48988.2020.9238097.
XUE Lan-qing, LI **ao-peng, ZHANG N L. Not all attention is needed: Gated attention network for sequence data [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34(4): 6550–6557. DOI: https://doi.org/10.1609/aaai.v34i04.6129.
WANG Zi-cheng, CHEN Li-ren, DING Zhen-ni, et al. An enhanced interval PM2.5 concentration forecasting model based on BEMD and MLPI with influencing factors [J]. Atmospheric Environment, 2020, 223: 117200. DOI: https://doi.org/10.1016/j.atmosenv.2019.117200.
ZHOU Min, WANG Bo, GUO Shu-dong, et al. Multi-objective prediction intervals for wind power forecast based on deep neural networks [J]. Information Sciences, 2021, 550: 207–220. DOI: https://doi.org/10.1016/j.ins.2020.10.034.
JIANG **, LI Chen, LI Ran-ran, et al. An innovative hybrid air pollution early-warning system based on pollutants forecasting and Extenics evaluation [J]. Knowledge-Based Systems, 2019, 164: 174–192. DOI: https://doi.org/10.1016/j.knosys.2018.10.036.
LIU Hui, DUAN Zhu, CHEN Chao. A hybrid framework for forecasting PM2.5 concentrations using multi-step deterministic and probabilistic strategy [J]. Air Quality, Atmosphere & Health, 2019, 12(7): 785–795. DOI: https://doi.org/10.1007/s11869-019-00695-8.
YONG N K, AWANG N. Wavelet-based time series model to improve the forecast accuracy of PM10 concentrations in Peninsular Malaysia [J]. Environmental Monitoring and Assessment, 2019, 191(2): 1–12. DOI: https://doi.org/10.1007/s10661-019-7209-6.
LIU Hui, DUAN Zhu, CHEN Chao. A hybrid multiresolution multi-objective ensemble model and its application for forecasting of daily PM2.5 concentrations [J]. Information Sciences, 2020, 516: 266–292. DOI: https://doi.org/10.1016/j.ins.2019.12.054.
SALINAS D, FLUNKERT V, GASTHAUS J, et al. DeepAR: Probabilistic forecasting with autoregressive recurrent networks [J]. International Journal of Forecasting, 2020, 36(3): 1181–1191. DOI: https://doi.org/10.1016/j.ijforecast.2019.07.001.
LI Guo-hui, CHEN Ling, YANG Hong. Prediction of PM2.5 concentration based on improved secondary decomposition and CSA-KELM [J]. Atmospheric Pollution Research, 2022, 13(7): 101455. DOI: https://doi.org/10.1016/j.apr.2022.101455.
PINSON P, TASTU J. Discussion of “prediction intervals for short-term wind farm generation forecasts” and “combined nonparametric prediction intervals for wind power generation” [J]. IEEE Transactions on Sustainable Energy, 2014, 5(3): 1019–1020. DOI: https://doi.org/10.1109/TSTE.2014.2323851.
NOWOTARSKI J, WERON R. Recent advances in electricity price forecasting: A review of probabilistic forecasting [J]. Renewable and Sustainable Energy Reviews, 2018, 81: 1548–1568. DOI: https://doi.org/10.1016/j.rser.2017.05.234.
HARR L, SINSEL T, SIMON H, et al. PM2.5 exposure differences between children and adults [J]. Urban Climate, 2022, 44: 101198. DOI: https://doi.org/10.1016/j.uclim.2022.101198.
XU Cheng-liang, CHEN Huan-xin. A hybrid data mining approach for anomaly detection and evaluation in residential buildings energy data [J]. Energy and Buildings, 2020, 215: 109864. DOI: https://doi.org/10.1016/j.enbuild.2020.109864.
JIANG Feng, ZHU Qian-nan, TIAN Tian-hai. An ensemble interval prediction model with change point detection and interval perturbation-based adjustment strategy: A case study of air quality [J]. Expert Systems With Applications, 2023, 222: 119823. DOI: https://doi.org/10.1016/j.eswa.2023.119823.
YAO Wei, ZENG Zhi-gang, LIAN Cheng. Generating probabilistic predictions using mean-variance estimation and echo state network [J]. Neurocomputing, 2017, 219(C): 536–547. DOI: https://doi.org/10.1016/j.neucom.2016.09.064.
Author information
Authors and Affiliations
Contributions
LIU Hui provided the concept and edited the draft of the manuscript. FANG Ya-min conducted the literature review and wrote the first draft of the manuscript. FANG Ya-min edited the draft of manuscript. All authors replied to reviewers’ comments and revised the final version.
Corresponding author
Ethics declarations
FANG Ya-min and LIU Hui declare that they have no conflict of interest.
Additional information
Foundation item: Project(2020YFC2008605) supported by the National Key Research and Development Program, China; Project (52072412) supported by the National Natural Science Foundation of China; Project(2021JJ30359) supported by the Natural Science Foundation of Hunan Province of China; Project(2020TJ-Q06) supported by the Hunan Province Science and Technology Talent Support Project, China
Rights and permissions
About this article
Cite this article
Fang, Ym., Liu, H. Probabilistic concentration prediction of PM2.5 in subway stations based on multi-resolution elastic-gated attention mechanism and Gaussian mixture model. J. Cent. South Univ. 30, 2818–2832 (2023). https://doi.org/10.1007/s11771-023-5401-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11771-023-5401-x