Hybrid VMD-CNN1D Framework: Evaluating Decomposition’s Contribution to PM2.5 Prediction

Dwi Yuwono; Arya Adhyaksa Waskita; Tukiyat Tukiyat

doi:10.30871/jaic.v10i4.13685

Authors

Dwi Yuwono Program Studi Teknik Informatika, Pascasarjana, Universitas Pamulang | Badan Meteorologi, Klimatologi, dan Geofisika
Arya Adhyaksa Waskita Program Studi Teknik Informatika, Pascasarjana, Universitas Pamulang
Tukiyat Tukiyat Program Studi Teknik Informatika, Pascasarjana, Universitas Pamulang

DOI:

https://doi.org/10.30871/jaic.v10i4.13685

Keywords:

Air Quality, Bias Correction, Convolutional Neural Network, PM2.5 Prediction, Variational Mode Decomposition

Abstract

Fine particulate matter (PM2.5) pollution in Jakarta reached an annual average of 37.3 µg/m³ in 2023, 7.4 times the WHO threshold, causing over 10,000 premature deaths annually. Accurate short-term prediction is essential for early-warning systems, yet two gaps remain common in decomposition-based deep learning literature: Variational Mode Decomposition (VMD) parameters are rarely selected via systematic sensitivity analysis, and prediction bias is rarely corrected explicitly. This study addresses both gaps using 32,144 PM2.5 observations from Jakarta (2015-2025). A grid search over 20 parameter combinations identified K=6, alpha=500 as optimal (reconstruction error 1.88%). Each of six IMFs was predicted independently using CNN1D, reconstructed additively, and bias-corrected using validation-set mean bias error. Decomposition, not model complexity, drove accuracy: a non-decomposed baseline reached only R²=0.4302, versus R²=0.9151 (RMSE=4.83 µg/m³) for the proposed framework. We further validated fixed- and adaptive-parameter variants (VMD, AVMD) across 5 independent runs. VMD-CNN1D achieved R²=0.9177±0.0173, RMSE=4.7346±0.5064, while AVMD-CNN1D (K=7 selected consistently) achieved R²=0.9238±0.0089, RMSE=4.5388±0.2649. Diebold-Mariano tests showed AVMD outperforming VMD in 4 of 5 runs (p<0.0001) with lower variance, though a paired t-test across runs was not significant (p=0.684). Ablation identified the lowest-frequency IMF as most critical, consistent with Jakarta's dry-season and land-fire pollution patterns, while residual analysis revealed heteroscedasticity as a limitation. These findings show that rigorous parameter selection, bias correction, and multi-run validation, not architectural complexity alone, make decomposition-based deep learning reliable for PM2.5 prediction, offering a reproducibility-aware baseline for a future Jakarta air-quality early-warning system.

Downloads

Download data is not yet available.

References

[1] IQAir, “2023 World Air Quality Report,” 2024.

[2] WHO, “WHO global air quality guidelines Particulate matter (PM2.5 and PM10), ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide,” 2021.

[3] K. Dragomiretskiy and D. Zosso, “Variational mode decomposition,” IEEE Transactions on Signal Processing, vol. 62, no. 3, pp. 531–544, Feb. 2014, doi: 10.1109/TSP.2013.2288675.

[4] Huang, S. R. L, H. H. Shih, N. Y. i-C, and H. H. and, N. E. L, “The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis,” 1998.

[5] Z. Wu and N. E. Huang, “Ensemble Empirical Mode Decomposition: A Noise-Assisted Data Analysis Method,” 2009.

[6] A. Khattak, S. Alotaibi, R. N. Alahmadi, C. M. Matara, and S. Taglawi, “Hybrid VMD–BiGRU Framework for Multi-Step Forecasting of PM2.5 in Traffic-Intensive Cities of the Kingdom of Saudi Arabia,” Atmosphere (Basel)., vol. 16, no. 12, Dec. 2025, doi: 10.3390/atmos16121324.

[7] X. Wu, J. Zhu, and Q. Wen, “Short-term prediction of PM2.5 concentration by hybrid neural network based on sequence decomposition,” PLoS One, vol. 19, no. 5 May, May 2024, doi: 10.1371/journal.pone.0299603.

[8] X. H. Wang et al., “Air quality forecasting using a spatiotemporal hybrid deep learning model based on VMD–GAT–BiLSTM,” Sci. Rep., vol. 14, no. 1, Dec. 2024, doi: 10.1038/s41598-024-68874-x.

[9] T. Zeng, L. Xu, Y. Liu, R. Liu, Y. Luo, and Y. Xi, “A hybrid optimization prediction model for PM2.5 based on VMD and deep learning,” Atmos. Pollut. Res., vol. 15, no. 7, Jul. 2024, doi: 10.1016/j.apr.2024.102152.

[10] T. Zeng et al., “A Deep Learning PM2.5 Hybrid Prediction Model Based on Clustering–Secondary Decomposition Strategy,” Electronics (Switzerland), vol. 13, no. 21, Nov. 2024, doi: 10.3390/electronics13214242.

[11] X. Bai, N. Zhang, X. Cao, and W. Chen, “Prediction of PM2.5 concentration based on a CNN-LSTM neural network algorithm,” PeerJ, vol. 12, no. 8, 2024, doi: 10.7717/peerj.17811.

[12] J. Zhang, X. Xin, Y. Shang, Y. Wang, and L. Zhang, “Nonstationary significant wave height forecasting with a hybrid VMD-CNN model,” Ocean Engineering, vol. 285, Oct. 2023, doi: 10.1016/j.oceaneng.2023.115338.

[13] S. Zhou, W. Wang, L. Zhu, Q. Qiao, and Y. Kang, “Deep-learning architecture for PM2.5 concentration prediction: A review,” Sep. 01, 2024, Editorial Board, Research of Environmental Sciences. doi: 10.1016/j.ese.2024.100400.

[14] C. J. Willmott, “On The Validation Of Models,” Phys. Geogr., vol. 2, no. 2, pp. 184–194, Jul. 1981, doi: 10.1080/02723646.1981.10642213.

[15] S. Ioffe and C. Szegedy, “Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift,” Mar. 2015, [Online]. Available: http://arxiv.org/abs/1502.03167

[16] N. Srivastava, G. Hinton, A. Krizhevsky, and R. Salakhutdinov, “Dropout: A Simple Way to Prevent Neural Networks from Overfitting,” 2014.

[17] D. P. Kingma and J. Ba, “Adam: A Method for Stochastic Optimization,” Jan. 2017, [Online]. Available: http://arxiv.org/abs/1412.6980

[18] Q. Yang, Q. Yuan, T. Li, H. Shen, and L. Zhang, “The relationships between PM2.5 and meteorological factors in China: Seasonal and regional variations,” Int. J. Environ. Res. Public Health, vol. 14, no. 12, Dec. 2017, doi: 10.3390/ijerph14121510.

[19] M. A. A.-A. A. S. Hashim, “Air Quality and Particulate Matter Forecasting using VMD-BiGRU Deep Learning Model in Saudi Arabia Cities,” Atmosphere (Basel)., vol. 16, no. 1, pp. 245–263, 2025.

[20] Z. Zhang and S. Zhang, “Modeling air quality PM2.5 forecasting using deep sparse attention-based transformer networks,” International Journal of Environmental Science and Technology, vol. 20, no. 12, pp. 13535–13550, Dec. 2023, doi: 10.1007/s13762-023-04900-1.