基于ARIMA模型的浙江省GDP预测模型研究

Research on the GDP Forecasting Model of Zhejiang Province Based on ARIMA Model

摘要: 通过选取1952年至2020年浙江省国内生产总值的相关数据，建立了时间序列分析中的自回归滑动平均求和模型ARIMA(p,d,q)，利用该模型对浙江省GDP进行短期预测，为浙江省经济的发展提供参考。建立1952年至2018年浙江省GDP数据的时间序列，利用R语言软件建立ARIMA模型，并用该模型预测的2019年和2020年浙江省GDP数据与实际数据进行比较，对建立的模型进行优化评估，最后利用优化模型对2021年至2023年浙江省GDP进行短期预测。根据建立的时间序列分析得到最优模型为ARIMA(4,2,3)，预测值与实际值的平均相对误差为2.28%，ARIMA模型能较好地反映浙江省GDP发展的趋势并进行短期预测。

Abstract: By selecting the relevant data of the GDP of Zhejiang Province from 1952 to 2020, the autoregres-sive moving average summation model ARIMA(p,d,q) in the time series analysis is established. The short-term forecast provides a reference for the economic development of Zhejiang Province. Estab-lish a time series of Zhejiang Province’s GDP data from 1952 to 2018, use R language software to establish an ARIMA model, and use the model to compare the Zhejiang Province’s GDP data in 2019 and 2020 with the actual data to optimize and evaluate the established model. Finally, the optimi-zation model is used to make a short-term forecast of Zhejiang Province’s GDP from 2021 to 2023. According to the established time series analysis, the optimal model is ARIMA(4,2,3), and the aver-age relative error between the predicted value and the actual value is 2.28%. The ARIMA model can better reflect the trend of GDP development in Zhejiang Province and make short-term predictions.