Comparative Analysis of ARMA and ARIMA Models for Air Quality Prediction

Accurate prediction of urban air pollution is crucial for environmental management and public health. In this study, two classical time series models were comparatively analyzed to simulate and predict the concentrations of PM2.5, NO2, and O3 in Shanghai from 2019 to 2021. Based on the Interquartile Range method, we performed data preprocessing. The AutoRegressive Moving Average and the AutoRegressive Integrated Moving Average models were applied to predict the air pollutants. We used various evaluation metrics to compare the model performance. The conclusion shows that ARIMA is more suitable for air pollution applications as it captures the nonlinear trend of the data through differential methods.