Mazuruse, Gideon, Nyagadza, Brighton ORCID: https://orcid.org/0000-0001-7226-0635, Tumbure, Akinson, Makoni, Tendai and Muvuti, Ashley (2025) Algorithmic Optimization for Efficient Air Quality Prediction Models through Machine Learning: A Case Study of Shillong City in India. Next Research (100346). p. 100346. (In Press)

Full text not available from this repository.

Abstract

Poor air quality is a severe problem, potentially causing health and environmental problems. The study investigated the effectiveness of machine learning methods in air quality prediction, using the datasets from Shillong City in India. Key variables used in this study were particulate matter (PM2.5 and PM10), carbon monoxide (CO), nitrogen dioxide (NO2), nitric oxide (NO), sulphur dioxide (SO2), ammonia (NH3), and ozone (O3). The machine learning methods used in this study were AdaBoost, LightGBM and Support Vector Machine. Correlation analysis of variables with the Air Quality Index (AQI) revealed that O3 and NO had weak correlations with the AQI (r = 0.37 and 0.01, respectively), which justified their exclusion from model construction. The dataset was divided into training (70%) and testing (30%) subsets. LightGBM outperformed other models, achieving an accuracy of 0.86, whilst the AdaBoost had the second-best results. The Support Vector Machine had an accuracy of 0.827. This illustrates the usefulness of the LightGBM method in predicting air quality, providing knowledge for efficient air quality management. PM2.5 and PM10 had the greatest impact on LightGBM model results, and this illustrates the usefulness of monitoring these two air pollutants in air quality management. The LightGBM model proved to be very efficacious in predicting air quality in the given environment, considering the complex nature of the data. The model results may inform interventions for air pollution management in similar areas. Policies and regulations should therefore pay greater attention to sources of particulate matter pollution in Shillong City and develop appropriate interventions.

Item Type:	Article
Status:	In Press
DOI:	10.1016/j.nexres.2025.100346
School/Department:	London Campus
URI:	https://ray.yorksj.ac.uk/id/eprint/11929

University Staff: Request a correction | RaY Editors: Update this record