摘要
针对时间序列的Hurst指数的估计方法的问题,目前国内外已经提出了R/S,DAF,绝对值法,周期图法等多种方法。但上述方法都会对Hurst指数的估计值产生易误解和不一致的结果。针对这个问题,通过对R/S分析法,小波分析法,迭代估计算法和Whittle法的描述,进行数值模拟来说明这些方法所得Hurst指数估计值的误差,通过比较能够得出Whittle法是一种具有更高精度和更好稳定性的方法。
At present, there are many methods of analyzing the Hurst exponent of time series in and abroad, but most of them could lead to misunderstanding and disagreement to the estimation of Hurst exponent. This paper describes R/S analysis, wavelet analysis, iterative algorithm estimation and modified Whittle analysis, by the means of numerical simulation, to compare the error of estimation of Hurst exponent that obtained by the above methods. The comparison is made to test and verify Whittle method has the best accuracy and stability.
出处
《计算机工程与应用》
CSCD
2014年第16期154-158,共5页
Computer Engineering and Applications
