Estimation Of Air Quality Index In Delhi By Merging Neural Networks And Multiple Regression Techniques with Principal Components Analysis

Abstract

A major focus of air quality research in recent years has been the AQI measurement as a way to gauge the harm pollution does to people's health and well-being in cities. Air Quality Index (AQI) accuracy is the primary goal of this research, which uses PCA and Artificial Neural Network (ANN) approaches. Our investigators are looking into the city of Delhi. To forecast the air quality index, The main components score (PCS) of 11 historical air quality and meteorological data is used in an ANN model (AQI). Delhi is the subject of this investigation. In order to make accurate forecasts of the air quality index, ANN models make use of the main components score (PCS) of 11 meteorological and historical air quality indicators (AQI). A comparison is made between ANN and MLR models, which are commonly used to estimate the AQI. Other than PCA, you may also reduce the eleven parameters to just eight PCs. PC-ANN (PC-ANN) models use the eight PCs as input data. The R2, RMSE, MAPE, and MAE values were used to make comparisons between the various models and hypotheses. The PC-ANN model outperforms all others when considering the complexities of air pollution. As a result, the PC-ANN approach may be utilized to make better decisions and address atmospheric management challenges.