我国经济增长预测研究新探——基于分类变量的扩散指标法

Chinese Economic Growth Rate Forecasting:Based on Classification of Diffusion Index

基于分类变量的扩散指标方法是在Stock and Watson所提出的扩散指标法的基础上,利用变量之间的相似系数先对变量进行聚类分析,根据各变量的相似程度的大小分为几个变量子集,然后分别从变量子集中提取扩散指标。应用基于变量分类的扩散指标方法对我国经济增长的预测结果表明,同扩散指标法相比,该方法更加充分利用了原始数据中的有用信息,预测精度很高,因而是对经济增长进行预测的另一种有效方法。

The classification of diffusion index is based on diffusion index which is proposed by Stock and Watson. Its use the similarity coefficient of the variable to clustering analysis, then according the size of variables to divide into several subsets of variables, the last we extract diffusion index from subsets. Compared with diffusion index, the classification of diffusion index is more fully utilize the raw data and make more accuracy in forecasting Chinese economic growth rate, so it＇s the economic growth rate forecasting for another effective way.