Pollutant Significance Assessment and AQI Forecasting Framework: Based on Linear Regression and xLSTM

Air pollution has become a global environmental problem worth considering, so accurate prediction and targeted plans for air pollution are growing increasingly important. This study aims to predict the Air Quality Index (AQI) using extended Long Short-Term Memory (xLSTM) on an air quality dataset spanning 5 years, leveraging its enhanced sequence modeling capabilities., in order to provide a scientific basis for environmental policy formulation and public health intervention. Also, linear regression was used to study which specific air pollutants have the greatest impact on AQI. This enables the environment-related institutions to develop targeted plans to control air pollution. The result of the study suggests that CO and PM2.5 are the pollutants worth most concern. Moreover, the accuracy of xLSTM in predicting AQI values that belong to the same category as the actual AQI values is 79.29%. Collectively, this integrated predictive approach provides a valuable scientific foundation for evidence-based environmental policy formulation and public health intervention.