摘要
由于区域经济系统中许多经济变量呈现出强非线性与大波动性的特征,使得传统的时间序列线性建模和预测技术难以适应区域经济预测的要求.为此,提出基于支持向量机改进的残差自回归区域经济预测模型.首先采用时间序列分析中的残差自回归模型对时间序列趋势进行线性拟合,然后对残差自回归模型估计后的残差序列采用支持向量回归方法再次提取其非线性特征,从而提高区域经济时间序列模型的预测精度.最后以广东省GDP的预测实例说明模型的有效性.
For many economic variables in regional economy system have features of non- linearity and instability, the result of prediction, achieved by traditional linear modeling and predicting technology, doesn＇t meet the demand of accuracy. Thus, a regioni economic predic- tion model based on Residual Auto-regressive method improved by Support Vector Regression method is proposed in this paper. First, we fit the linear part of time series using Residual Auto-regressive method, then draws the nonlinear part from the residual sequence using Sup- port Vector Regression method. The new model helps to increase the accuracy of prediction in regional economy system. At the end, a prediction of GDP in Guangdong province shows the affectivity of the model.
出处
《数学的实践与认识》
CSCD
北大核心
2013年第19期36-42,共7页
Mathematics in Practice and Theory
基金
广东省自然科学基金(S201101006103)
江门市2012年度产业技术研究与开发项目
关键词
区域经济
预测
支持向量回归
时间序列分析
regional economic
prediction
support vector regression
time series analyze
