Faced with the rapid update of nonlinear and irregular big data from the environmental monitoring system, both the public and managers urgently need reliable methods to predict possible air pollutions in the future. Therefore, a multi-scale deep learning (MDL) and optimal combination ensemble (OCE) approach for hourly air quality index (AQI) forecasting is proposed in this paper, named MDL-OCE model. Before normal modeling, all original data are preprocessed through missing data filling and outlier testing to ensure smooth computation. Due to the complexity of such big data, slope-based ensemble empirical mode decomposition (EEMD) is adopted to decompose the time series of AQI and meteorological conditions into a finite number of simple intrinsic mode function (IMF) components and one residue component. Then, to unify the number of components of different variables, the fine-to-coarse (FC) technique is used to reconstruct all components into high frequency component (HF), low frequency component (LF), and trend component (TC). For purpose of extracting the underlying relationship between AQI and meteorological conditions, the three components are respectively trained and predicted by different deep learning architectures (stacked sparse autoencoder (SSAE)) with a multilayer perceptron (MLP). The corresponding forecasting results of three components are merged by OCE method to better achieve the ultimate AQI forecasting outputs. The empirical results clearly demonstrate that our proposed MDL-OCE model outperforms other advanced benchmark models in terms of forecasting performances in all cases.
ZhaiB. and ChenJ., Development of a stacked ensemble model for forecasting and analyzing daily average PM2.5 concentrations in Beijing, China, Sci Total Environ635 (2018), 644–658.
2.
ApteJ.S., MarshallJ.D., CohenA.J. and BrauerM., Addressing global mortality from ambient PM2.5, Environ Sci Technol49(13) (2015), 8057–8066.
3.
LelieveldJ., EvansJ.S., FnaisM., GiannadakiD. and PozzerA., The contribution of outdoor air pollution sources to premature mortality on a global scale, Nature525(7569) (2015), 367–371.
4.
ZhuJ., WuP., ChenH., ZhouL. and TaoZ., A hybrid forecasting approach to air quality time series based on endpoint condition and combined forecasting model, Int J Environ Res Public Health15 (2018), 1941–1960.
5.
KumarA., PatilR.S., DikshitA.K., IslamS. and KumarR., Evaluation of control strategies for industrial air pollution sources using American meteorological society/environmental protection agency regulatory model with simulated meteorology by weather research and forecasting model, J Clean Prod116 (2016), 110–117.
6.
ChenD., LiuX., LangJ., ZhouY., WeiL., WangX. and GouX., Estimating the contribution of regional transport to PM2.5 air pollution in a rural area on the North China Plain, Sci Total Environ583 (2017), 280–291.
7.
JacksonB., ChauD., GurerK., et al., Comparison of ozone simulations using MM5 and CALMET/MM5 hybrid meteorological fields for the July/August CCOS episode, Atmos Environ40 (2006), 2812–2822.
8.
ChengX., LiuY., XuX., et al., Lidar data assimilation method based on CRTM and WRF-Chem models and its application in PM2.5 forecasts in Beijing, Sci Total Environ682 (2019), 541–552.
9.
ReyesJ.M., XuY., VizueteW., et al., Regionalized PM2.5 Community Multiscale Air Quality model performance evaluation across a continuous spatiotemporal domain, Atmos Environ148 (2017), 258–265.
10.
DimitriouK. and KassomenosP., A study on the reconstitution of daily PM10 and PM2.5 levels in Paris with a multivariate linear regression model, Atmos Environ98 (2014), 648–654.
11.
BasM.D.C., OrtizJ., BallesterosL. and MartorellS., Evaluation of a multiple linear regression model and SARIMA model in forecasting 7Be air concentrations, Chemosphere177 (2017), 326–333.
12.
NebenzalA. and FishbainB., Long-term forecasting of nitrogen dioxide ambient levels in metropolitan areas using the discrete-time Markov model, Environ Modell Softw107 (2018), 175–185.
13.
DingS., DangY., LiX., WangJ. and ZhaoK., Forecasting Chinese CO2 emissions from fuel combustion using a novel grey multivariable model, J Clean Prod162 (2017), 1527–1538.
14.
SinghV., Higher pollution episode detection using image classification techniques, Environ Model Assess21(5) (2016), 591–601.
15.
NeaguC., AvourisN., KalapanidasE., et al., Neural and Neuro-Fuzzy Integration in a Knowledge-Based System for Air Quality Prediction, Appl Intell17 (2002), 141–169.
16.
LiY. and TaoY., Daily PM10 concentration forecasting based on multiscale fusion support vector regression, J Intell Fuzzy Syst34(6) (2018), 3833–3844.
17.
AbderrahimH., ChellaliM.R. and HamouA., Forecasting PM10 in Algiers: efficacy of multilayer perceptron networks, Environ Sci Pollut Res23 (2016), 1634–1641.
18.
WangZ., ChenL., ZhuJ., ChenH. and YuanH., Double decomposition and optimal combination ensemble learning approach for interval-valued AQI forecasting using streaming data, Environ Sci Pollut Res27 (2020), 37802–37817.
19.
SucarL.E., Pérez-BritoJ., Ruiz-SuárezJ.C., et al., Learning structure from data and its application to ozone prediction, Appl Intell7 (1997), 327–338.
20.
GuoL., ChenB., ZhangH., et al., A new approach combining a simplified FLEXPART model and a Bayesian-RAT method for forecasting PM10 and PM2.5, Environ Sci Pollut Res27 (2020), 2165–2183.
21.
WangD., WeiS., LuoH., YueC. and GrunderO., A novel hybrid model for air quality index forecasting based on two-phase decomposition technique and modified extreme learning machine, Sci Total Environ580 (2017), 719–733.
22.
YuanW., WangK., BoX., TangL. and WuJ., A novel multi-factor & multi-scale method for PM2.5 concentration forecasting, Environ Pollut255 (2019), 113187.
23.
HintonG.E., OsinderoS. and TehY.W., A fast learning algorithm for deep belief nets, Neural Comput18(7) (2006), 1527–1554.
24.
XieJ., WangX., LiuY. and BaiY., Autoencoder-based deep belief regression network for air particulate matter concentration forecasting, J Intell Fuzzy Syst34(6) (2018), 3475–3486.
25.
WenC., LiuS., YaoX., et al., A novel spatiotemporal convolutional long short-term neural network for air pollution prediction, Sci Total Environ654 (2019), 1091–1099.
26.
BaiY., LiY., ZengB., LiC. and ZhangJ., Hourly PM2.5 concentration forecast using stacked autoencoder model with emphasis on seasonality, J Clean Prod224 (2019), 739–750.
27.
LiH., WangJ., LiR. and LuH., Novel analysis–forecast system based on multi-objective optimization for air quality index, J Clean Prod208(20) (2019), 1365–1383.
28.
WangJ., DuP., HaoY., MaX., NiuT. and YangW., An innovative hybrid model based on outlier detection and correction algorithm and heuristic intelligent optimization algorithm for daily air quality index forecasting, J Environ Manage255 (2020), 109855.
29.
NiX., HuangH. and DuW., Relevance analysis and short-term prediction of PM2.5 concentrations in Beijing based on multi-source data, Atmos Environ150 (2017), 146–161.
30.
YangQ., YuanQ., LiT., ShenH. and ZhangL., The Relationships between PM2.5 and Meteorological Factors in China: Seasonal and Regional Variations, Int J Environ Res Public Health14 (2017), 1510.
31.
LiuW., GuoG., ChenF. and ChenY., Meteorological pattern analysis assisted daily PM2.5 grades prediction using SVM optimized by PSO algorithm, Atmos Pollut Res10 (2019), 1482–1491.
32.
JianL., ZhaoY., ZhuY.P., ZhangM.B. and BertolattiD., An application of ARIMA model to predict submicron particle concentrations from meteorological factors at a busy roadside in Hangzhou, China, Sci Total Environ426(2) (2012), 336–345.
33.
LiX., PengL., YaoX., CuiS., HuY., YouC. and ChiT., Long short-term memory neural network for air pollutant concentration predictions: method development and evaluation, Environ Pollut231 (2017), 997–1004.
34.
ZhangL., LiuY. and ZhaoF., Important meteorological variables for statistical long-term air quality prediction in eastern China, Theor Appl Climatol134(1-2) (2018), 25–36.
35.
BaiY., LiY., WangX., XieJ. and LiC., Air pollutants concentrations forecasting using back propagation neural network based on wavelet decomposition with meteorological conditions, Atmos Pollut Res7 (2016), 557–566.
36.
WuQ. and LinH., A novel optimal-hybrid model for daily air quality index prediction considering air pollutant factors, Sci Total Environ683 (2019), 808–821.
37.
WangZ., ChenL., DingZ. and ChenH., An enhanced interval PM2.5 concentration forecasting model based on BEMD and MLPI with influencing factors, Atmos Environ223 (2020), 117200.
38.
ChenS., WangJ. and ZhangH., A hybrid PSO-SVM model based on clustering algorithm for short-term atmospheric pollutant concentration forecasting, Technol Forecast Soc Chang146 (2019), 41–54.
39.
HuangN.E., ShenZ., LongS.R., et al., The Empirical Mode Decomposition and Hilbert spectrum for nonlinear and non-stationary time series analysis, Proc Roy Soc London A454 (1998), 903–995.
40.
WuZ. and HuangN.E., Ensemble empirical mode decomposition: a noise-assisted data analysis method, Advances in Adaptive Data Analysis1(1) (2009), 1–41.
41.
XiongT., BaoY. and HuZ., Does restraining end effect matter in EMD-based modeling framework for time series prediction? Some experimental evidences, Neurocomputing123 (2014), 174–184.
42.
DätigM. and SchlurmannT., Performance and limitations of the Hilbert–Huang transformation (HHT) with an application to irregular water waves, Ocean Engineering31 (2004), 1783–1834.
43.
ZhangX., LaiK.K. and WangS.Y., A new approach for crude oil price analysis based on empirical mode decomposition, Energy Economics30 (2008), 905–918.
44.
ZhuB., ShiX., ChevallierJ., et al., An adaptive multiscale ensemble learning paradigm for nonstationary and nonlinear energy price time series forecasting, Journal of Forecasting35(7) (2016), 633–651.
45.
BengioY., CourvilleA. and VincentP., Unsupervised Feature Learning and Deep Learning: A Review and New Perspectives. Cornell University Library arXiv (2014): 1206.5538v3.
46.
LvY., DuanY., KangW., et al., Traffic flow prediction with big data: a deep learning approach, IEEE Trans Intell Transp Syst16(2) (2015), 865–873.
47.
GuoZ., ZhaoW., LuH. and WangJ., Multi-step forecasting for wind speed using a modified EMD-based artificial neural network model, Renewable Energy37(1) (2012), 241–249.
48.
YangH.F. and ChenY.P., Hybrid deep learning and empirical mode decomposition model for time series applications, Expert Syst Appl120 (2019), 128–138.
49.
LiuZ., JiangP., ZhangL. and NiuX., A combined forecasting model for time series: Application to short-term wind speed forecasting, Appl Energy259 (2020), 114137.
50.
DieboldF.X. and MarianoR.S., Comparing predictive accuracy, J Bus Econ Stat13(3) (1995), 134–144.
51.
WangJ., AthanasopoulosG., HyndmanR.J. and WangS., Crude oil price forecasting based on internet concern using an extreme learning machine, Int J Forecast34 (2018), 665–677.
