Optimal Weighted Combination Forecasting Based on Holt-Winters and SARIMA Intervention Models

Abstract

Excessive emissions of SO₂ will lead to the formation of acid rain. Accurate prediction of the content of SO₂ in the atmosphere can provide reasonable suggestions for the government's policy formulation and residents' travel. Each prediction model has its drawbacks. For complex problems such as weather prediction, using a single prediction model often leads to a relatively large error. Optimal weighted prediction is a commonly used model combination prediction method, which usually yields better prediction results. In order to accurately predict the content of SO₂ in the atmosphere of Urumqi City, the SARIMA intervention model, the Holt-Winters additive model, and the optimal weighted combination model composed when the weight coefficients of the two are  and  respectively were used to predict the content of SO₂ in the atmosphere of Urumqi City from April 2024 to March 2025. And the fitting situations and prediction accuracies of the three models were compared. As a result, the  values of these three models are ,  and  respectively, the  values are ,  and  respectively, and the  values are ,  and  respectively. The prediction effect of the combined model is the best. Finally, the content of SO₂ in the atmosphere of Urumqi City is predicted.