Impact of air pollutants on climate change and prediction of air quality index using machine learning models

• Published in: Environmental research 2023-12, Vol.239, p.117354-117354, Article 117354
• Main Authors: Ravindiran, Gokulan, Rajamanickam, Sivarethinamohan, Kanagarathinam, Karthick, Hayder, Gasim, Janardhan, Gorti, Arunkumar, Priya, Arunachalam, Sivakumar, AlObaid, Abeer A., Warad, Ismail, Muniasamy, Senthil Kumar
• Format: Article
• Language: English
• Subjects: Air pollution; Air quality index; Climate action; Machine learning
• ISSN: 0013-9351; 1096-0953
• DOI: 10.1016/j.envres.2023.117354

The impact of air pollution in Chennai metropolitan city, a southern Indian coastal city was examined to predict the Air Quality Index (AQI). Regular monitoring and prediction of the Air Quality Index (AQI) are critical for combating air pollution. The current study created machine learning models such as XGBoost, Random Forest, BaggingRegressor, and LGBMRegressor for the prediction of the AQI using the historical data available from 2017 to 2022. According to historical data, the AQI is highest in January, with a mean value of 104.6 g/gm, and the lowest in August, with a mean AQI value of 63.87 g/gm. Particulate matter, gaseous pollutants, and meteorological parameters were used to predict AQI, and the heat map generated showed that of all the parameters, PM2.5 has the greatest impact on AQI, with a value of 0.91. The log transformation method is used to normalize datasets and determine skewness and kurtosis. The XGBoost model demonstrated strong performance, achieving an R2 (correlation coefficient) of 0.9935, a mean absolute error (MAE) of 0.02, a mean square error (MSE) of 0.001, and a root mean square error (RMSE) of 0.04. In comparison, the LightGBM model's prediction was less effective, as it attained an R2 of 0.9748. According to the study, the AQI in Chennai has been increasing over the last two years, and if the same conditions persist, the city's air pollution will worsen in the future. Furthermore, accurate future air quality level predictions can be made using historical data and advanced machine learning algorithms. [Display omitted] •Predicted AQI for the city Chennai, South India.•PM2.5, Gaseous Pollutants and Metrological Parameters used to predict AQI.•XGboost model outperformed other models With an R2 value of 0.9935.•National wide lockdown due to Covid19 resulted in reduced AQI in the year 2020.