【摘要】 首先采用一个分形整合模型——误差逗留模型 (Error-Duration Model)仔细推导了分形时间序列过程的性质 ,特别是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 ;然后以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更多还原

【Abstract】 This article uses a model, Error\|Duration Model, to infer the characters of time\|sequence\|procedure, especially the characters of the sequence′s self\|relevant coefficient, and indicates the procedure has large difference with regular time\|sequence analysis tools. The article also give an example of a real time\|sequence to explain that procedure has the better accuracy than the traditional analysis tools in economic forecasting.更多还原