摘要
运用EMD分解和ARMA模型相结合的方法对福建省固定资产投资增长速度进行了预测。首先利用EMD技术对原始信号进行分解,得到不同频率的平稳的本征模函数,然后以各征模函数建立ARMA模型,最后将各征模函数的预测值进行合成,实现对原始数据序列的预测,提高了预测精度。结果显示:2000-2017年福建省固定资产投资增速预测,EMD-ARMA模型的平均预测误差为5.022 42%,比单一ARMA模型的平均预测误差6.023 48%减小16.6%。基于EMD-ARMA模型的预测,得到2018年福建省固定资产投资增长速度为13.890 28%。
The growth rate of fixed assets investment in Fujian province is forecasted by combining EMD decomposition with ARMA model.First, the EMD technology is used to decompose the original signal, and the stationary eigenmode function of different frequencies is obtained. Then, the ARMA model is established with each eigenmode function. Finally, the predictive value of each model function is synthesized to realize the prediction of the original data sequence and improve the prediction accuracy.The results show that the average prediction error of the EMD-ARMA model is 5.022 42%, which is 16.6% lower than the average prediction error of the single ARMA model of 6.023 48%. From the prediction of the model,the growth rate of fixed assets investment in Fujian in 2018 was 13.890 28%.
作者
舒服华
SHU Fuhua(School of Mechanical and Electrical Engineering,Wuhan University of Technology, Hubei 430070 ,China)
出处
《泉州师范学院学报》
2018年第5期77-81,108,共6页
Journal of Quanzhou Normal University
