固定资产投资与经济增长的非参数回归预测分析

分析了我国全社会固定资产投资与经济增长之间的关系。由于两者的长期均衡关系受到政治经济因素影响存在两个突变点,而带有虚拟变量的线性回归模型不能理想地捕获两者之间的动态关系。通过运用Local Linear非参数回归理论,对新中国成立以来的国内生产总值与固定资产投资的关系,分别建立了分段与整段非参数回归预测模型以及参数线性回归预测模型,并加以比较,结果表明:整段非参数回归模型拟合较好,并且能够做出高精度的预测。

To Analyse on Forecasting of the Nonparametric Regression for Fixed Investments between Economy Growing

The relationship between total investments in fixed assets and GDP of China is analyzed in this paper. The traditional linear regression model cannot ideally catch the dynamic characteristic of their relationship. So the nonparametric regression forecast model of GDP was made based on local linear nonparametric estimation theory. The result showed that the nonparametric regression model could achieve the higher precision including fitting and forecast data than the linear model of GDP.