云南进出口贸易额时序分析与预测

Time Series Analysis and Prediction of Import and Export Trade Volume of Yunnan Province

摘要: 本文基于1980年至2023年云南省进出口贸易年度数据，构建了经典的时间序列ARIMA模型进行分析与预测。研究结果显示单一的ARIMA模型在拟合云南进出口贸易数据趋势方面表现不佳，且预测精度较差。有鉴于此，本文通过引入外生干预事件，构建干预分析模型对云南进出口贸易数据进行分析与预测。研究表明，干预分析模型可以较好地拟合云南进出口贸易数据的发展趋势，且干预分析模型的预测性能相较于ARIMA模型有大幅改善。

Abstract: This study analyzes Yunnan Province’s import-export trade volume data from 1980 to 2023 employing the classical time-series ARIMA model for analysis and forecasting. The results show that the standalone ARIMA model performs inadequately in fitting the trends of Yunnan’s import and export trade data and exhibits poor prediction accuracy. In light of this, the study introduces exogenous intervention events to construct an intervention analysis model for analyzing and forecasting Yunnan’s import and export trade volume data. The research demonstrates that the intervention analysis model can effectively capture the developmental trends of Yunnan’s import-export volume data, and its predictive performance shows significant improvement compared to the ARIMA model.