# 基于时间序列分析的福建省GDP预测研究Forecast and Research on GDP of Fujian Province Based on Time Series Analysis

DOI: 10.12677/SA.2020.93046, PDF, HTML, XML, 下载: 73  浏览: 1,656

Abstract: Firstly, based on the time series analysis theory, the GDP of Fujian province from 1978 to 2015 was fitted and analyzed by Eviews software; the ARIMA(4,2,2) model and ARIMA(4,3,2) model were established. Secondly, using the two models, the GDP of Fujian province in 2016~2018 was used to test. We found the prediction error of ARIMA(4,3,2) is smaller, so it is considered that the model has a good effect. Finally using ARIMA(4,3,2) model, the GDP of Fujian province in 2019~2023 was predicted.

1. 引言

2. 数据收集

Table 1. GDP of Fujian province from 1978 to 2018

3. ARIMA(4,2,2)模型的建立

3.1. 平稳性检验

Figure 1. Sequence diagram of Fujian province’s GDP from 1978~2015

Figure 2. Unit root test of {X}

Figure 3. Unit root test of {LNX}

Figure 4. Unit root test of {D(LNX)}

Figure 5. Unit root test of {D(LNX,2)}

3.2. 模型识别

Figure 6. Autocorrelation diagram and partial correlation diagram of {D(LNX,2)}

3.3. 参数估计

Figure 7. Modeling result of ARIMA(4,2,2) after the AR(1), AR(3), MA(1) are eliminated

${x}_{t}=-0.0025-1.155004{x}_{t-2}-0.493897{x}_{t-4}+{\epsilon }_{t}-0.999939{\epsilon }_{t-2}$

3.4. 模型检验

ARIMA(4,2,2)模型残差的自相关系数均在二倍标准差内，P值也均大于0.005，所以可以认为序列之间没有任何关联，即为白噪声序列，可以用该模型进行预测。

4. ARIMA(4,3,2)模型的建立

4.1. 平稳性检验

Figure 8. Residual correlation diagram of ARIMA(4,2,2) model

Figure 9. Unit root test of {D(LNX,3)}

4.2. 模型识别

4.3. 参数估计

${x}_{t}=0.004544-1.272193{x}_{t-2}-0.605696{x}_{t-4}+{\epsilon }_{t}-0.999971{\epsilon }_{t-2}$

4.4. 模型检验

Figure 10. Autocorrelation diagram and partial correlation diagram of {D(LNX,3)}

Figure 11. Modeling result of ARIMA(4,3,2) after the AR(1), AR(3), MA(1) are eliminated

ARIMA(4,3,2)模型残差的自相关系数均在二倍标准差内，P值也均大于0.005，所以可以认为该模型序列之间没有任何关联，即为白噪声序列，所以可以用于预测。

5. 预测结果检验

Figure 12. Residual correlation diagram of ARIMA(4,3,2) model

Table 2. Comparison of model results

Table 3. GDP forecast of Fujian province

6. 总结与展望

 [1] 宋喆磊. 改革开放以来福建省经济增长因素的实证分析[J]. 企业导报, 2009(6) : 112-113. [2] 李守丽. 时间序列模型在地级市GDP预测中的应用[D]: [硕士学位论文]. 郑州: 郑州大学, 2013. [3] 戴琳琳. 基于ARIMA模型的青岛市GDP预测分析[J]. 河北能源职业技术学院学报, 2019(3): 60-62. [4] 张旭昌. 基于时间序列分析的河北省GDP预测模型研究[J]. 改革与开放, 2018(21): 16-18. [5] 王鄂, 张霆. 时间序列在湖南省GDP预测中的应用[J]. 青岛大学学报(自然科学版), 2019, 32(3): 136-140. [6] 孙铁轩, 邵春福. 基于ARIMA与信息粒化SVR组合模型的交通事故时序预测[J]. 清华大学学报(自然科学版), 2014, 54(3): 348-359. [7] 福建统计局官网. 1978-2018年我国生产总值[EB/OL]. http://tjj.fujian.gov.cn/, 2020-02-01.