Prediction of PM_2.5 levels using ANN and MLR machine learning models for Lucknow in India

As industrialization has grown in recent years, air pollution has become a serious issue across the entire globe. It will only get worse in the future, especially in countries like India and China. As a result, controlling air pollution is essential for protecting human lives while maintaining the balance of industrial activity. The main aim of the research is to figure out the extent to which airborne pollution influences the inhabitants of big cities. This will help us estimate the air quality index (AQI), a measure that provides a way to assess air quality at a specific location by taking into account the total impact of key airborne contaminants. In the current research, the Air Quality Index in the city of Lucknow, Uttar Pradesh, India, is calculated by applying the Multiple Linear Regression (MLR) and Artificial Neural Network (ANN) models. The main aim is to evaluate the differences between the progression time, error operations, number of epochs, correlation factor, and predicted accuracy of three different neural network functions - Tansig-Tansig, Tansig-Purelin, and Purelin-Purelin - as well as the adaptation functions learngd and learngdm. Neural network training functions on the air pollution data points allow officials in estimating AQI for metropolitan areas through strategic decisions. The data from the central public school area monitoring station from 2017 to 2021 is used for the analysis using MATLAB. Six pollutants (PM2.5, SO2, NOx, CO, O3, and NH3) are examined in order to determine the AQI. Statistical analysis is performed on both ANN and MLR models. ANN is found to be capturing the air pollution behavior reasonably accurately for AQI forecasting.