Review on Air Pollution Forecasting Using Machine Learning on Time Series Data

Main Article Content

Samriddhi Sharma
Prof. Anjali Saxena

Abstract

Forecasting air pollution is vital for the purposes of air quality monitoring, air pollution control and sustainable management of air quality. This review intends to present and compare different machine learning and deep learning methods and hybrid approaches for forecasting air pollution concentration from time-series environmental data, focusing on the most commonly-used pollutants, meteorological parameters, datasets, preprocessing methods, forecasting strategies, and evaluation aspects. A systematic review of literature was carried out by analyzing different conventional machine learning models like Linear Regression, Random Forest, Support Vector Regression, Gradient Boosting, and XGBoost and deep-learning models including Artificial Neural Networks, LSTM, GRU, Bidirectional GRU, and hybrid CNN based models. In addition, various data preparation techniques such as missing value treatment, feature engineering, normalization and splitting of time-series were explored. The reviewed studies showed that traditional machine-learning models are computationally efficient, whereas deep-learning models have more powerful abilities of learning nonlinear relations and temporal dependency. The hybrid models additionally incorporate feature extraction and temporal learning of complex multivariate environmental data. There are still challenges in data quality control, model interpretability, computational costs, real-time implementation, model comparison and prediction under sudden changes of pollution. Strong models with interpretability that are also computationally efficient and can estimate longer forecasting periods and multivariate environmental data are needed for future research.

Article Details

How to Cite
Samriddhi Sharma, & Prof. Anjali Saxena. (2026). Review on Air Pollution Forecasting Using Machine Learning on Time Series Data . International Journal of Advanced Research and Multidisciplinary Trends (IJARMT), 3(3), 1270–1282. Retrieved from https://ijarmt.com/index.php/j/article/view/1355
Section
Articles

References

A. Dairi, F. Harrou, S. Khadraoui, and Y. Sun, “Integrated Multiple Directed Attention-based Deep Integrated Multiple Directed Attention-based Deep Learning for Improved Air Pollution

Forecasting,” 2026, doi: 10.1109/tim.2021.3091511.

P. Linardatos, V. Papastefanopoulos, T. Panagiotakopoulos, and S. Kotsiantis, “OPEN CO 2 concentration forecasting in smart cities using a hybrid ARIMA – TFT model on multivariate time series IoT data,” Sci. Rep., no. 0123456789, pp. 1–22, 2026, doi: 10.1038/s41598-02342346-0.

C. Wu, R. Wang, S. Lu, J. Tian, L. Yin, and L. Wang, “Time-Series Data-Driven PM 2. 5 Forecasting: From Theoretical Framework to Empirical Analysis,” no. 2, 2025.

Y. Özüpak and F. Alpsalaz, “Air Quality Forecasting Using Machine Learning: Comparative Analysis and Ensemble Strategies for Enhanced Prediction,” Water, Air, Soil Pollut., vol. 236, no. 7, pp. 1–17, 2025, doi: 10.1007/s11270-025-08122-8.

T. Septi, H. Irie, A. Dimara, and K. Wikantika, “Machine learning-based global air quality index development using remote sensing and ground-based stations,” Environ. Adv., vol. 15, no. August 2023, p. 100456, 2024, doi: 10.1016/j.envadv.2023.100456.