摘要
熵估计在生理时间序列上被广泛应用,例如近似熵与样本熵。将熵估计方法应用于汇率时间序列中,用于识别汇率市场在不同时间的波动状态并加以分析。在不同维度下讨论了近似熵与样本熵反映汇率时间序列波动的情况,并对同一维度下近似熵与样本熵效果进行了比较,发现样本熵比近似熵可以更好地反映汇率的波动性,并且更加灵敏。得出样本熵算法在汇率市场中良好地反映了大事件的发生与延续,解决了近似熵算法对微小的复杂性变化不灵敏的缺陷,并且显著提高了熵估计在短时间序列上的可用性和精确度。
Entropy estimation is widely used in physiological time series,such as the approximate entropy and the sample entropy. In this paper,entropy estimation method is used to identify and analyze the time series of fluctuations in different times in the exchange rate market. This paper discussed the approximate entropy and the sample entropy on the reflects of the fluctuation of the exchange rate time series under different dimensions,and compared the approximate entropy and the sample entropy under the same dimension,and found that the sample entropy can reflect the volatility of exchange rate better than the approximate entropy,and more sensitive. The paper found that the sample entropy algorithm reflects the big events and continuance in the currency markets very well,solves the defect that the approximate entropy algorithm is not sensitive to the tiny change of complexity,and significantly improves the availability and accuracy of the entropy estimation in a short time series.
出处
《技术与创新管理》
2016年第4期438-441,共4页
Technology and Innovation Management
关键词
汇率
波动性
近似熵
样本熵
exchange rate
volatility
approximate entropy
sample entropy
