分形整合过程在经济预测中的应用

The application of fractional integration process to economic forecasting

本文首先采用一个分形整合模型──误差逗留模型（Ｅｒｒｏｒ－Ｄｕｒａｔｉｏｎ Ｍｏｄｅｌ）仔细推导了一类具有分形时间序列过程的性质，特别是其序列自相关系数的性质，表明分形整合过程与常规的时间序列分析工具有很大的不同，然后就一个实际的时间序列为例，说明了分形整合过程在经济预测中的应用比传统的分析工具有较好的预测精度。

The paper used a fractional integration model—an error-duration model to infer the characteris- tics of a kind of fractional time series processes and the properties of its auto correlation coefficient in par- ticular, 'which showed that the fractional integration process is different greatly from the analysing too1 of normal time series. The paper, then, specified that the application of the fractional integration process is more acucrate than the traditiona1 analysing tool in economic forecasting by a real example of time series.