Enhanced Air Quality Prediction Using Optimized Sarimax Model for Coimbatore Region

This research delves into examining the forecast of air quality in Coimbatore, India, according to data gleaned from the Central Pollution Control Board (CPCB) specific to the area. The gathered information underwent a preprocessing step, using interpolation to ensure uniformity and completeness. Forecasting methods such as Seasonal Autoregressive Integrated Moving Average with Exogenous Regressors (SARIMAX) and Long Short-Term Memory (LSTM) models were utilized to predict air quality levels. The SARIMAX model was fine-tuned through a grid search process, while the LSTM underwent hyperparameter adjustments to improve air quality predictions. Performance evaluation of these models involved metrics like R-squared, Mean absolute error, and Root mean square error. Based on the selected evaluation criteria, the outcome shows that SARIMAX outperformed LSTM in Coimbatore when it came to predicting the air quality. This analysis improves our understanding of air quality patterns in Coimbatore. Highlights SARIMAX's effectiveness as a modeling tool for atmospheric science applications.