能源需求非线性组合预测中包容性检验方法的改进

Improved Encompassing Test in Energy Demand Nonlinear Combined Forecasting Model

传统包容性检验方法不能很好地适应能源需求非线性组合模型的特征,基于非线性函数的思想,本文提出将传统包容性检验的线性回归模型改进为非线性回归模型,利用逐步删除单一预测模型的组合模型与原组合模型之间的非线性包容性检验方法,确定该单一模型是否参与非线性组合预测。实证结果显示,改进的包容性检验方法在能源需求非线性组合模型的适用性及预测精度方面均有所提高。

Traditional encompassing test can＇t adapt to the characteristics of energy demand nonlinear combined model. Based on the principle of nonlinear function, this paper improves the traditional linear regression model into nonlinear regression model in encompassing tests, and used nonlinear encompassing test between the original combined model and the combined model which is by step-wise deletion of single forecasting model, in order to determine whether the single model is involved in the nonlinear combined forecasting. The empirical results show that the improved encompassing test get higher applicability and precision in the field of energy demand nonlinear combined model.