Comparative Analysis of Air Quality Forecasting Using Machine and Deep Learning Based Approaches

Many emerging nations are experiencing severe air pollution on account of the growing urbanization and industry. Because air pollution has an impact on global human health and sustainable development, governments and it is becoming source of concern in society. Anticipating air pollution thus became...

Full description

Saved in:
Bibliographic Details
Published in2025 International Conference on Artificial Intelligence and Data Engineering (AIDE) pp. 8 - 12
Main Authors HK, Sowmya, Chandresh, Koyal, J., Jesy Janet Kumari, D'Souza, Sonia Maria
Format Conference Proceeding
LanguageEnglish
Published IEEE 06.02.2025
Subjects
Air pollution
Air Quality Index
Analytical models
Atmospheric modeling
Computational modeling
Deep learning
Indexes
Long short term memory
Long Short-Term Memory (LSTM)
Machine Learning
Nitrogen
Predictive models
Urban planning
Online AccessGet full text
DOI10.1109/AIDE64228.2025.10987543

Cover

More Information
Summary:Many emerging nations are experiencing severe air pollution on account of the growing urbanization and industry. Because air pollution has an impact on global human health and sustainable development, governments and it is becoming source of concern in society. Anticipating air pollution thus became essential. Prediction of air quality has gained popularity by virtue of the volume and a range of data that air quality surveillance systems have collected. This paper uses a Python framework to analyze the air quality, implementing historical air quality data analysis for Bangalore city and predicting its quality exercising a number of machine and deep learning (D-Learning) models. In the proposed work, 5 distinct Machine learning(M-Learning) models including Random Forest regression (RFR), Support Vector Regression (SVR), Linear regression (LR), Decision Tree (DT), and Long Short- Term Model of D-Learning Technique (D-LSTM) have been accustomed to compute the Air Quality Index (AQI) of Bangalore city. This strategy has been put into practice in Google Collab with GPU run time. As stated by the measures including Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE), D-Learning- based LSTM model has demonstrated better reliability than other M-Learning models. Performance metrics values for LSTM is superior compared to the rest of M-Learning models
DOI:10.1109/AIDE64228.2025.10987543