PM_2.5 probabilistic forecasting system based on graph generative network with graph U-nets architecture

Abstract

Urban air pollution has brought great troubles to physical and mental health, economic development, environmental protection, and other aspects. Predicting the changes and trends of air pollution can provide a scientific basis for governance and prevention efforts. In this paper, we propose an interval prediction method that considers the spatio-temporal characteristic information of PM_2.5 signals from multiple stations. K-nearest neighbor (KNN) algorithm interpolates the lost signals in the process of collection, transmission, and storage to ensure the continuity of data. Graph generative network (GGN) is used to process time-series meteorological data with complex structures. The graph U-Nets framework is introduced into the GGN model to enhance its controllability to the graph generation process, which is beneficial to improve the efficiency and robustness of the model. In addition, sparse Bayesian regression is incorporated to improve the dimensional disaster defect of traditional kernel density estimation (KDE) interval prediction. With the support of sparse strategy, sparse Bayesian regression kernel density estimation (SBR-KDE) is very efficient in processing high-dimensional large-scale data. The PM_2.5 data of spring, summer, autumn, and winter from 34 air quality monitoring sites in Beijing verified the accuracy, generalization, and superiority of the proposed model in interval prediction.

Keywords

PM_2.5 interval forecasting / graph generative network / graph U-Nets / sparse Bayesian regression / kernel density estimation / spatial-temporal characteristics

References

[1] Wang Z-C, Gao R-B, Wang P, et al. . A new perspective on air quality index time series forecasting: A ternary interval decomposition ensemble learning paradigm [J]. Technological Forecasting and Social Change. 2023, 191: 122504

[2] Zhao X-J, Zhang Z-Y, Xu J, et al. . Impacts of aerosol direct effects on PM_2.5 and O₃ respond to the reductions of different primary emissions in Beijing-Tianjin-Hebei and surrounding area [J]. Atmospheric Environment. 2023, 309: 119948

[3] Rakholia R, Le Q, Vu K, et al. . AI-based air quality PM_2.5 forecasting models for developing countries: A case study of Ho Chi Minh City, Vietnam [J]. Urban Climate. 2022, 46: 101315

[4] Adani M, D’Isidoro M, Mircea M, et al. . Evaluation of air quality forecasting system FORAIR-IT over Europe and Italy at high resolution for year 2017 [J]. Atmospheric Pollution Research. 2022, 13(6): 101456

[5] Wang X-L, Xie N-M, Yang L. A flexible grey Fourier model based on integral matching for forecasting seasonal PM_2.5 time series [J]. Chaos, Solitons & Fractals. 2022, 162: 112417

[6] Li Y-Z, Guo J-E, Sun S-L, et al. . Air quality forecasting with artificial intelligence techniques: A scientometric and content analysis [J]. Environmental Modelling & Software. 2022, 149: 105329

[7] Aggarwal A, Toshniwal D. A hybrid deep learning framework for urban air quality forecasting [J]. Journal of Cleaner Production. 2021, 329: 129660

[8] Liu H, Yang R. A spatial multi-resolution multi-objective data-driven ensemble model for multi-step air quality index forecasting based on real-time decomposition [J]. Computers in Industry. 2021, 125: 103387

[9] Elbaz K, Hoteit I, Shaban W M, et al. . Spatiotemporal air quality forecasting and health risk assessment over smart city of NEOM [J]. Chemosphere. 2023, 313: 137636

[10] Li H-M, Wang J-Z, Yang H-F, et al. . Air quality deterministic and probabilistic forecasting system based on hesitant fuzzy sets and nonlinear robust outlier correction [J]. Knowledge-Based Systems. 2022, 237: 107789

[11] Gao X, Li W-D. A graph-based LSTM model for PM_2.5 forecasting [J]. Atmospheric Pollution Research. 2021, 12(9): 101150

[12] Zhan H-L, Zhu X, Hu J-M. A probabilistic forecasting approach for air quality spatio-temporal data based on kernel learning method [J]. Applied Soft Computing. 2023, 132: 109858

[13] Shen J, Valagolam D, Mccalla S. Prophet forecasting model: A machine learning approach to predict the concentration of air pollutants (PM_2.5, PM₁₀, O₃, NO₂, SO₂, CO) in Seoul, South Korea [J]. Peer J. 2020, 8: e9961

[14] Lu X, Zhou W, Qi C-C, et al. . Prediction into the future: A novel intelligent approach for PM_2.5 forecasting in the ambient air of open-pit mining [J]. Atmospheric Pollution Research. 2021, 12(6): 101084

[15] Wang J-Z, Wang R, Li Z-W. A combined forecasting system based on multi-objective optimization and feature extraction strategy for hourly PM_2.5 concentration [J]. Applied Soft Computing. 2022, 114: 108034

[16] Kim B Y, Lim Y K, Cha J W. Short-term prediction of particulate matter (PM₁₀ and PM_2.5) in Seoul, South Korea using tree-based machine learning algorithms [J]. Atmospheric Pollution Research. 2022, 13(10): 101547

[17] Erkin N, Simayi M, Ablat X, et al. . Predicting spatiotemporal variations of PM_2.5 concentrations during spring festival for county-level cities in China using VIIRS-DNB data [J]. Atmospheric Environment. 2023, 294: 119484

[18] Samal K K R, Babu K S, Das S K. Multi-directional temporal convolutional artificial neural network for PM_2.5 forecasting with missing values: A deep learning approach [J]. Urban Climate. 2021, 36: 100800

[19] Hao X-T, Hu X-J, Liu T, et al. . Estimating urban PM_2.5 concentration: An analysis on the nonlinear effects of explanatory variables based on gradient boosted regression tree [J]. Urban Climate. 2022, 44: 101172

[20] Liu H, Zhang X-Y, Yang Y-X, et al. . Hourly traffic flow forecasting using a new hybrid modelling method [J]. Journal of Central South University. 2022, 29(4): 1389-1402

[21] Yang R, Liu H, Li Y-F. A heterogeneous ensemble architecture coupling model selection sorting and residual error iterative correction for crude oil price forecasting [J]. Applied Soft Computing. 2023, 148: 110865

[22] Kong Y-W, Sheng L-F, Li Y-P, et al. . Improving PM_2.5 forecast during haze episodes over China based on a coupled 4D-LETKF and WRF-Chem system [J]. Atmospheric Research. 2021, 249: 105366

[23] Cheng F-Y, Feng C Y, Yang Z M, et al. . Evaluation of real-time PM_2.5 forecasts with the WRF-CMAQ modeling system and weather-pattern-dependent bias-adjusted PM_2.5 forecasts in Taiwan [J]. Atmospheric Environment. 2021, 244: 117909

[24] Liu H, Yang R, Duan Z. Wind speed forecasting using a new multi-factor fusion and multi-resolution ensemble model with real-time decomposition and adaptive error correction [J]. Energy Conversion and Management. 2020, 217: 112995

[25] Liu H, Yang R, Duan Z, et al. . A hybrid neural network model for marine dissolved oxygen concentrations time-series forecasting based on multi-factor analysis and a multi-model ensemble [J]. Engineering. 2021, 7(12): 1751-1765

[26] Li Y-F, Liu Z-Y, Liu H. A novel ensemble reinforcement learning gated unit model for daily PM_2.5 forecasting [J]. Air Quality, Atmosphere & Health. 2021, 14(3): 443-453

[27] Zhao L-X, Li Z-Y, Qu L-L. Forecasting of Beijing PM_2.5 with a hybrid ARIMA model based on integrated AIC and improved GS fixed-order methods and seasonal decomposition [J]. Heliyon. 2022, 8(12): e12239

[28] Janarthanan R, Partheeban P, Somasundaram K, et al. . A deep learning approach for prediction of air quality index in a metropolitan city [J]. Sustainable Cities and Society. 2021, 67: 102720

[29] Alruwaili O, Kostanic I, Al-Sabbagh A, et al. . IoT based: Air quality index and traffic volume correlation [C]. 2020 11th IEEE Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON). 2020, New York, NY, USA, IEEE: 143-147

[30] Yang R, Liu H, Nikitas N, et al. . Short-term wind speed forecasting using deep reinforcement learning with improved multiple error correction approach [J]. Energy. 2022, 239: 122128

[31] Liu H, Yang R, Wang T-T, et al. . A hybrid neural network model for short-term wind speed forecasting based on decomposition, multi-learner ensemble, and adaptive multiple error corrections [J]. Renewable Energy. 2021, 165: 573-594

[32] Yang R, Liu H, Li Y-F. An ensemble self-learning framework combined with dynamic model selection and divide-conquer strategies for carbon emissions trading price forecasting [J]. Chaos, Solitons & Fractals. 2023, 173: 113692

[33] Zhou Y-L, Chang F J, Chen H, et al. . Exploring Copula-based Bayesian model averaging with multiple ANNs for PM_2.5 ensemble forecasts [J]. Journal of Cleaner Production. 2020, 263: 121528

[34] Liu H, Deng D-H. An enhanced hybrid ensemble deep learning approach for forecasting daily PM_2.5 [J]. Journal of Central South University. 2022, 29(6): 2074-2083

[35] Huang J-H, Liu H. A hybrid decomposition-boosting model for short-term multi-step solar radiation forecasting with NARX neural network [J]. Journal of Central South University. 2021, 28(2): 507-526

[36] Prasad K, Gorai A K, Goyal P. Development of ANFIS models for air quality forecasting and input optimization for reducing the computational cost and time [J]. Atmospheric Environment. 2016, 128: 246-262

[37] Chang Y-S, Chiao H T, Abimannan S, et al. . An LSTM-based aggregated model for air pollution forecasting [J]. Atmospheric Pollution Research. 2020, 11(8): 1451-1463

[38] Menares C, Perez P, Parraguez S, et al. . Forecasting PM_2.5 levels in Santiago de Chile using deep learning neural networks [J]. Urban Climate. 2021, 38: 100906

[39] Eren B, Aksangür İ, Erden C. Predicting next hour fine particulate matter (PM_2.5) in the Istanbul Metropolitan City using deep learning algorithms with time windowing strategy [J]. Urban Climate. 2023, 48: 101418

[40] Tan J, Liu H, Li Y-F, et al. . A new ensemble spatiotemporal PM_2.5 prediction method based on graph attention recursive networks and reinforcement learning [J]. Chaos, Solitons & Fractals. 2022, 162: 112405

[41] Yang R, Liu H, Li Y-F. Quantifying uncertainty of marine water quality forecasts for environmental management using a dynamic multi-factor analysis and multiresolution ensemble approach [J]. Chemosphere. 2023, 331: 138831

[42] Mandal S, Thakur M. A city-based PM_2.5 forecasting framework using spatially attentive cluster-based graph neural network model [J]. Journal of Cleaner Production. 2023, 405: 137036

[43] Qi Y-L, Li Q, Karimian H, et al. . A hybrid model for spatiotemporal forecasting of PM_2.5 based on graph convolutional neural network and long short-term memory [J]. Science of the Total Environment. 2019, 664: 1-10

[44] Dun A, Yang Y-N, Lei F. Dynamic graph convolution neural network based on spatial-temporal correlation for air quality prediction [J]. Ecological Informatics. 2022, 70: 101736

[45] Wang Z-C, Chen L-R, Ding Z-N, et al. . An enhanced interval PM_2.5 concentration forecasting model based on BEMD and MLPI with influencing factors [J]. Atmospheric Environment. 2020, 223: 117200

[46] Wang Z-C, Li H, Chen H-Y, et al. . Linear and nonlinear framework for interval-valued PM_2.5 concentration forecasting based on multi-factor interval division strategy and bivariate empirical mode decomposition [J]. Expert Systems with Applications. 2022, 205: 117707

[47] Jiang L-Y, Tao Z-F, Zhu J-M, et al. . Exploiting PSO-SVM and sample entropy in BEMD for the prediction of interval-valued time series and its application to daily PM_2.5 concentration forecasting [J]. Applied Intelligence. 2023, 53(7): 7599-7613

[48] Liu H, Duan Z, Chen C. A hybrid framework for forecasting PM_2.5 concentrations using multi-step deterministic and probabilistic strategy [J]. Air Quality, Atmosphere & Health. 2019, 12(7): 785-795

[49] Wang Z-C, Chen L-R, Zhu J-M, et al. . Double decomposition and optimal combination ensemble learning approach for interval-valued AQI forecasting using streaming data [J]. Environmental Science and Pollution Research. 2020, 27(30): 37802-37817

[50] Jadhav A, Pramod D, Ramanathan K. Comparison of performance of data imputation methods for numeric dataset [J]. Applied Artificial Intelligence. 2019, 33(10): 913-933

[51] Beretta L, Santaniello A. Nearest neighbor imputation algorithms: A critical evaluation [J]. BMC Medical Informatics and Decision Making. 2016, 16(Suppl 3): 74

[52] Li C-X, Welling M, Zhu J, et al. . Graphical generative adversarial networks [C]. Proceedings of the 32nd International Conference on Neural Information Processing System. 2018, Montreal, Canada, Curran Associates Inc: 6072-6083

[53] Zhou D-W, Zheng L-C, Han J-W, et al. . A data-driven graph generative model for temporal interaction networks [C]. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 2020401-411

[54] Du Y-Q, Guo X-J, Cao H-N, et al. . Disentangled spatiotemporal graph generative models [J]. Proceedings of the AAAI Conference on Artificial Intelligence. 2022, 36(6): 6541-6549

[55] Gao H-Y, Ji S-W. Graph U-nets [C]. Proceedings of the 36th International Conference on Machine Learning. 2019, Long Beach, California, PMLR: 20832092

[56] Yan Z-Y, Cao W-M, Ji J-H. Social behavior prediction with graph U-Net+ [J]. Discover Internet of Things. 2021, 1(1): 18

[57] Liu H, Duan Z, Chen C. A hybrid multiresolution multi-objective ensemble model and its application for forecasting of daily PM_2.5 concentrations [J]. Information Sciences. 2020, 516: 266-292

[58] Yang S-M, Wu H-Y. A novel PM_2.5 concentrations probability density prediction model combines the least absolute shrinkage and selection operator with quantile regression [J]. Environmental Science and Pollution Research. 2022, 29(52): 78265-78291

RIGHTS & PERMISSIONS

Central South University