1. MMC
  2. All Volumes and Issues
  3. Volume 10, Number 4, 2023
  4. A hybrid model for predicting air quality combining Holt–Winters and Deep Learning Approaches: A novel method to identify ozone concentration peaks

A hybrid model for predicting air quality combining Holt–Winters and Deep Learning Approaches: A novel method to identify ozone concentration peaks

MMC.
2023;
Volume 10, Number 4
: pp. 1154–1163
https://doi.org/10.23939/mmc2023.04.1154
Received: August 20, 2023
Accepted: October 23, 2023

Mathematical Modeling and Computing, Vol. 10, No. 4, pp. 1154–1163 (2023)

Authors:
  1. N. Marrakchi
  2. A. Bergam
  3. H. Fakhouri
  4. K. Kenza
1
SMAD, FPL, Abdelmalek Essaadi University
2
SMAD, FPL, Abdelmalek Essaadi University
3
SMAD, FPL, Abdelmalek Essaadi University
4
DGM, National Climate Center, Air Quality Department, General Directorate of Meteorology

Ozone (O$_3$) from the troposphere is one of the substances that has a strong effect on air pollution in the city of Tanger.  Prediction of this pollutant can have positive improvements in air quality.  This paper presents a new approach combining deep-learning algorithms and the Holt–Winters method in order to detect pollutant peaks and obtain a more accurate forecasting model.  Given that LSTM is an extremely powerful algorithm, we hybridized with the Holt–Winters method to enhance the model.  Making use of multiple accuracy metrics, the models' efficiency is investigated.  Empirical findings reveal the superiority of the hybrid model by providing forecasts that are more accurate with an index of agreement equal to 0.91.

Air quality forecasting
Ozone (O3)
long short-term memory (LSTM)
Holt-Winters method
Recurrent Neural Network (RNN)
artificial neural networks
  1. Samadi A., Achelhi H.  Industry 4.0 in The Economic Activity Zones in Morocco: Tangier-Tetouan-Alhoceima Region Case.  International Journal of Accounting, Finance, Auditing, Management and Economics.  2 (6-1), 327–338 (2021).
  2. Zhang B., Song C., Li Y., Jiang X.  Spatiotemporal prediction of $O_3$ concentration based on the KNN-Prophet-LSTM model.  Heliyon.  8 (11), e11670 (2022).
  3. Lim C. C., Hayes R. B., Ahn J., Shao Y., Silverman D. T., Jones R. R., Garcia C., Bell M. L., Thurston G. D.  Long-term exposure to ozone and cause-specific mortality risk in the united states.  American Journal of Respiratory and Critical Care Medicine.  200 (8), 1022–1031 (2019).
  4. Suraboyina S., Allu S. K., Anupoju G. R., Polumati A.  A comparative predictive analysis of back-propagation artificial neural networks and non-linear regression models in forecasting seasonal ozone concentrations.  Journal of Earth System Science.  131 (3), 189 (2022).
  5. Kovač-Andrić E., Sheta A., Faris H., Gajdošik M. Š.  Forecasting ozone concentrations in the east of Croatia using nonparametric Neural Network Models.  Journal of Earth System Science.  125, 997–1006 (2016).
  6. Ensafi Y., Amin S. H., Zhang G., Shah B.  Time-series forecasting of seasonal items sales using machine learning–a comparative analysis.  International Journal of Information Management Data Insights.  2 (1), 100058 (2022).
  7. Chattopadhyay G., Chattopadhyay S.  Autoregressive forecast of monthly total ozone concentration: A neurocomputing approach.  Computers & Geosciences.  35 (9), 1925–1932 (2009).
  8. Akbarzadeh A., Vesali Naseh M., NodeFarahani M.  Carbon monoxide prediction in the atmosphere of tehran using developed support vector machine.  Pollution.  6 (1), 43–57 (2020).
  9. Kaur J., Parmar K. S., Singh S.  Autoregressive models in environmental forecasting time series: a theoretical and application review.  Environmental Science and Pollution Research.  30, 19617–19641 (2023).
  10. Oufdou H., Bellanger L., Bergam A., El Ghaziri A., Khomsi K., Qannari E. M., et al. Comparison of Different Regularized and Shrinkage Regression Methods to Predict Daily Tropospheric Ozone Concentration in the Grand Casablanca Area.  Advances in Pure Mathematics.  8 (10), 793 (2018).
  11. Hong F., Ji C., Rao J., Chen C., Sun W.  Hourly ozone level prediction based on the characterization of its periodic behavior via deep learning.  Process Safety and Environmental Protection.  174, 28–38 (2023).
  12. Tsai C.-h., Chang L.-c., Chiang H.-c.  Forecasting of ozone episode days by cost-sensitive neural network methods.  Science of the Total Environment.  407 (6), 2124–2135 (2009).
  13. Tamas W. W., Notton G., Paoli C., Nivet M.-L., Voyant C.  Hybridization of air quality forecasting models using machine learning and clustering: An original approach to detect pollutant peaks.  Aerosol and Air Quality Research.  16 (2), 405–416 (2016).
  14. Belavadi S. V., Rajagopal S., Ranjani R., Mohan R.  Air quality forecasting using LSTM RNN and wireless sensor networks.  Procedia Computer Science.  170, 241–248 (2020).
  15. Cinar Y. G., Mirisaee H., Goswami P., Gaussier E., Aït-Bachir A.  Period-aware content attention RNNs for time series forecasting with missing values.  Neurocomputing.  312, 177–186 (2018).
  16. Braik M., Sheta A., Al-Hiary H.  Hybrid neural network models for forecasting ozone and particulate matter concentrations in the Republic of China.  Air Quality, Atmosphere & Health.  13, 839–851 (2020).
  17. Jamei M., Ali M., Malik A., Karbasi M., Sharma E., Yaseen Z. M.  Air quality monitoring based on chemical and meteorological drivers: Application of a novel data filtering-based hybridized deep learning model.  Journal of Cleaner Production.  374, 134011 (2022).
  18. Maia A. L. S., de Carvalho F. D. A. T.  Holt's exponential smoothing and neural network models for forecasting interval-valued time series.  International Journal of Forecasting.  27 (3), 740–759 (2011).
  19. Dantas T. M., Oliveira F. L. C., Repolho H. M. V.  Air transportation demand forecast through Bagging Holt Winters methods.  Journal of Air Transport Management.  59, 116–123 (2017).
  20. Dullah H., Ahmed A. N., Kumar P., Elshafie A.  Integrated nonlinear autoregressive neural network and Holt Winters exponential smoothing for river streaming flow forecasting at Aswan High.  Earth Science Informatics.  16 (1), 773–786 (2023).
  21. Hyndman R., Koehler A. B., Ord J. K., Snyder R. D.  Forecasting with Exponential Smoothing: The State Space Approach.  Springer Science & Business Media (2008).
  22. Programmer L.  Deep Learning: Recurrent Neural Networks in Python, LSTM, GRU, and more RNN machine learning architectures in Python and Theano (Machine Learning in Python) (2016).
  23. Willmott C. J., Robeson S. M., Matsuura K.  A refined index of model performance.  International Journal of Climatology.  32 (13), 2088–2094 (2012).