Performance analysis of classical time series and machine learning models for PM2.5 and PM10 forecasting: A case study at an air quality monitoring station in Udon Thani, Thailand

Abstract

Air pollution forecasting is a critical component of air quality management, especially in areas experiencing persistent particulate matter issues. However, in practice, the performance of forecasting models depends on local climatic conditions and human activities in each area, resulting in varying forecasting patterns based on spatial contexts. This research aimed to compare the performance between time-series methodologies and machine learning in forecasting hourly PM2.5 and PM10 concentrations, focusing on analysis at air quality monitoring stations in Udon Thani province, Thailand, a medium- sized urban area influenced by both local emissions and regional meteorological patterns. This study used hourly datasets collected from January 1, 2023, to November 30, 2024, comprising a total of 16,800 records obtained from the pollution control department’s monitoring station at Nong Phra Chak Park. Since missing values are frequently encountered in real- world data collection, a linear interpolation method was employed to handle these gaps and enhance the efficiency of the proposed methods. The tested models included three time-series methods- seasonal autoregressive integrated moving average (SARIMA), seasonal naïve, and Holt–Winters exponential smoothing (ETS)-and four machine learning methods: random forest (RF), support vector regression (SVR), artificial neural networks (ANN), and extreme gradient boosting (XGBoost). The study utilized 80% of the data for training and 20% for testing, with performance evaluated through MAE, RMSE, R², MASE, and sMAPE indicators. The findings reveal that while time-series models can effectively reflect basic seasonal structures, they possess limited capability in capturing short-term fluctuations and sudden pollution surges, with MAE ranging from 8.096 to 14.106. In contrast, machine learning models consistently demonstrated lower forecasting errors, particularly the support vector regression methodology, which provided the most accurate and stable performance for both PM2.5 (MAE = 2.4899) and PM10 (MAE = 3.7340). This demonstrates its efficiency in modeling nonlinear relationships and short-term dynamics under the environmental conditions of Udon Thani province. The experimental results suggest that forecasting models aligned with local characteristics can yield reliable predictions, which will benefit environmental agencies and policymakers in developing air quality surveillance systems and implementing mitigation measures that effectively reflect regional specificities.

Keywords

air pollution forecasting / PM2.5 and PM10 / time-series models / machine learning and short-term forecasting

References

[1] Leão MLP, Zhang L, da Silva Júnior FMR (2023) Effect of particulate matter (PM2.5 and PM10) on health indicators: Climate change scenarios in a Brazilian metropolis. Environ Geochem Hlth 45: 2229-2240. https://doi.org/10.1007/s10653-022-01331-8

[2] World Health Organization, (2021) WHO global air quality guidelines: Particulate matter (PM2.5 and PM10), ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide, Geneva: World Health Organization. Available from: https://www.who.int/publications/i/item/9789240034228.

[3] Rochana K, Wongprachan R (2025) Statistical model for air quality forecasting: A case study of dust particles no larger than 2.5 microns (PM2.5) in Chiang Mai Province. NKRAFA J Sci Technol 21: 242-258.

[4] Li X, Zhang Y, Wang J, et al. (2023) Long-term forecasting of PM2.5 and PM10 concentrations and analysis of influencing factors. Sustainability 16: 19. https://doi.org/10.3390/su16010019

[5] Chelani AB, Devotta S (2006) Air quality forecasting using a hybrid autoregressive and nonlinear model. Atmos Environ 40: 1774-1780. https://doi.org/10.1016/j.atmosenv.2005.11.019

[6] Gardner MW, Dorling SR (1998) Artificial neural networks (the multilayer perceptron)-a review of applications in the atmospheric sciences. Atmos Environ 32: 2627-2636. https://doi.org/10.1016/S1352-2310(97)00447-0

[7] Grivas G, Chaloulakou A (2006) Artificial neural network models for prediction of PM10 hourly concentrations, in the Greater Area of Athens, Greece. Atmo Environ 40: 1216-1229. https://doi.org/10.1016/j.atmosenv.2005.10.037

[8] Zhang Y, Bocquet M, Mallet V, et al. (2012) Real-time air quality forecasting, part I: History, techniques, and current status. Atmos Environ 60: 632-655. https://doi.org/10.1016/j.atmosenv.2012.06.031

[9] Polsena T, Jitkongchuen D, Thusaranon P (2023) Integrated rearrange processing of hybrid model with weighted values for PM2.5 forecasting. ICITEE 213-216. https://doi.org/10.1109/icitee59582.2023.10317678

[10] Chiang PW, Horng SJ (2021) Hybrid time-series framework for daily-based PM2.5 forecasting. IEEE Access 9: 104162-104176. https://doi.org/10.1109/ACCESS.2021.3099111

[11] Chen D, Xu T, Li Y, et al. (2015) A hybrid approach to forecast air quality during high-PM concentration pollution period. Aerosol Air Qual Res 15: 1325-1337. https://doi.org/10.4209/AAQR.2014.10.0253

[12] Zhang X, Rui X, Xia X, et al. (2015) A hybrid model for short-term air pollutant concentration forecasting, In: 2015 IEEE International Conference on Service Operations And Logistics, And Informatics (SOLI), IEEE, 171-175. https://doi.org/10.1109/SOLI.2015.7367614

[13] Thanavanich T, Yaibuates M, Duangtang P, et al. (2022) Hybrid algorithms based on historical accuracy for forecasting particulate matter concentrations. IAES Int J Artif Intell 11: 1297-1305. https://doi.org/10.11591/ijai.v11.i4.pp1297-1305

[14] Pollution Control Department, Thailand air quality monitoring data, Ministry of Natural Resources and Environment, Thailand, 2024. Available from: http://air4thai.pcd.go.th/.

[15] Smith RL (1986) Time series analysis in acid rain modeling: Evaluation of filling missing values by linear interpolation. Atmos Environ 20: 1941-1943. https://doi.org/10.1016/0004-6981(86)90335-5

[16] McElroy TS, Politis DN (2022) Optimal linear interpolation of multiple missing values. Stat Infer Stoch Pro 25: 471-483. https://doi.org/10.1007/s11203-022-09269-5

[17] St-Aubin P, Agard B (2022) Precision and reliability of forecasts performance metrics. Forecasting 4: 882-903. https://doi.org/10.3390/forecast4040048

[18] Quiroz-Flores JC, et al. (2023) Forecasting analysis using machine learning models and statistical evaluation metrics. Int J Eng Trend Technol 71: 39-45. https://doi.org/10.14445/22315381/IJETT-V71I2P205

[19] Liu D (2024) The prediction and analysis of global climate change based on SARIMA. Appl Comput Eng 40: 268-273. https://doi.org/10.54254/2755-2721/40/20230665

[20] Mahanta N, Talukdar R (2024) Forecasting of electricity consumption by seasonal autoregressive integrated moving average model in Assam, India. Int J Energy Econ Policy 14: 651-658. https://doi.org/10.32479/ijeep.16506

[21] Jiang X, Xu L, Cui Y (2018) Seasonal model and its application in short-term forecasting. Int Conf Appl Math 194-196. https://doi.org/10.2991/AMMSA-18.2018.39

[22] Li X, Petropoulos F, Kang Y (2023) Improving forecasting by subsampling seasonal time series. Int J Prod Res 61: 976-992. https://doi.org/10.1080/00207543.2021.2022800

[23] Pan R (2010) Holt-winters exponential smoothing, Wiley Encyclopedia of Operations Research and Management Science. https://doi.org/10.1002/9780470400531.EORMS0385

[24] Lima S, Gonçalves AM, Costa M (2019) Time series forecasting using Holt-Winters exponential smoothing: An application to economic data. AIP Conf Proc 2186: 090003. https://doi.org/10.1063/1.5137999

[25] Alotaibi E, Nassif N (2024) Artificial intelligence in environmental monitoring: In-depth analysis. DIAI 4: 87. https://doi.org/10.1007/s44163-024-00198-1

[26] Milutinović M (2024) Machine learning in environmental monitoring, Facta Universitatis, Series: Working and Living Environmental Protection, 21: 155-164. https://doi.org/10.22190/fuwlep241029014m

[27] Hsieh WW (2025) Machine learning in environmental and climate science: Overview and introduction, Oxford Research Encyclopedia of Climate Science. https://doi.org/10.1093/acrefore/9780190228620.013.952

[28] Shalu (2023) Environmental monitoring with machine learning. EPRA Int J Multidiscip Res 9: 208-212. https://doi.org/10.36713/epra13330

[29] Peng N (2023) Application of machine learning techniques in environmental governance: A review. Adv Eng Technol Res 7: 528. https://doi.org/10.56028/aetr.7.1.528.2023

[30] Liu D, Fan Z, Fu Q, et al. (2019) Random forest regression evaluation model of regional flood disaster resilience based on the whale optimization algorithm. J Clean Prod 250: 119468. https://doi.org/10.1016/j.jclepro.2019.119468

[31] Rani KU, Venkataramana K (2025) Improved random forest regression for prediction. Int J Sci Technol Eng 13: 2084-2089. https://doi.org/10.22214/ijraset.2025.67722

[32] Hernández N, Biscay RJ, Talavera I (2007) Support vector regression methods for functional data, In: Progress in Pattern Recognition, Image Analysis and Applications, New York: Springer, 564-573. https://doi.org/10.1007/978-3-540-76725-1_59

[33] Abaszade M, Effati S (2018) Stochastic support vector regression with probabilistic constraints. Appl Intell 48: 243-256. https://doi.org/10.1007/S10489-017-0964-6

[34] Sonu, Bhokal RP (2017) Study of artificial neural network. Int J Math Trend Technol 47: 253-259. https://doi.org/10.14445/22315373/IJMTT-V47P535

[35] Rose A (2024) How do artificial neural networks work. J Adv Sci Technol 20: 172-177. https://doi.org/10.29070/ttrkmm98

[36] Bergou EH, Diouane Y, Kungurtsev V (2020) Convergence and complexity analysis of a Levenberg-Marquardt algorithm for inverse problems. J Optimiz Theory App 185: 927-944. https://doi.org/10.1007/s10957-020-01666-1

[37] Sheridan RP, Wang WM, Liaw A, et al. (2016) Extreme gradient boosting as a method for quantitative structure-Activity relationships. J Chem Inf Model 56: 2353-2360. https://doi.org/10.1021/acs.jcim.6b00591

[38] Chen T, Guestrin C (2016) XGBoost: A scalable tree boosting system, In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, New York: Association for Computing Machinery, 785-794. https://doi.org/10.1145/2939672.2939785