基于时间依赖的改进样本熵分析股票时间序列
|
摘要
样本熵是一个度量时间序列复杂度的非线性方法,广泛应用于各领域。然而,研究表明熵值的大小并不总是和时间序列的复杂性相关。为了解决这个问题,提出了多尺度熵,用来度量不同尺度下的时间序列的复杂度。但是,考虑到这种方法并没有解决样本熵在度量时间序列复杂度的问题,提出了基于时间依赖的改进样本熵,并将其用在股票收盘价和成交量时间序列上,研究它们对应的复杂度关系。同时,结合多尺度的方法,衡量不同尺度下股票收盘价时间序列和成交量时间序列的复杂性。实验结果表明,从收盘价时间序列和成交量时间序列的复杂度变化上能够揭示一定的股票的发展规律。另外,收盘价序列在不同的尺度上能够保持一致性,而成交量序列在不同的尺度上熵值变化则有不同的趋势,且股票类型越接近,熵值变化曲线也越接近。
Sample entropy is a nonlinear method to measure the complexity of time series and widely applied in various fields. However,studies have shown that the entropy is not always related to the complexity of time series. To solve this problem,multi-scale entropy is proposed to measure the complexity of time series over different scales. However,considering that this method does not solve the problem of sample entropy in measuring the complexity of time series,a modified sample entropy based on time dependent is proposed and applied in the stock closing price and volume time series to study their corresponding complexity relations. At the same time,combined with multi-scale method,the complexity of closing time series and volume time series is measured over different scales. The experiment shows that the complexity of the closing price time series and volume time series can reveal a certain rule of stock development. In addition,the closing price sequence can maintain consistency on different scales,while the entropy value changes of the volume sequence at different scales have different trends,and the closer the stock type is,the closer the entropy change curve is.
Sample entropy is a nonlinear method to measure the complexity of time series and widely applied in various fields. However,studies have shown that the entropy is not always related to the complexity of time series. To solve this problem,multi-scale entropy is proposed to measure the complexity of time series over different scales. However,considering that this method does not solve the problem of sample entropy in measuring the complexity of time series,a modified sample entropy based on time dependent is proposed and applied in the stock closing price and volume time series to study their corresponding complexity relations. At the same time,combined with multi-scale method,the complexity of closing time series and volume time series is measured over different scales. The experiment shows that the complexity of the closing price time series and volume time series can reveal a certain rule of stock development. In addition,the closing price sequence can maintain consistency on different scales,while the entropy value changes of the volume sequence at different scales have different trends,and the closer the stock type is,the closer the entropy change curve is.
引文
[1] 王怡.上市公司的网络舆情事件演化与股价变动的关联性研究[D].南京:南京理工大学,2017.
[2] 闫树熙.上证股市价格与成交量关系的实证研究[J].河南科学,2017,35(2):340-344.
[3] XU Mengjia,SHANG Pengjian,HUANG Jingjing.Modified generalized sample entropy and surrogate data analysis for stock markets[J].Communications in Nonlinear Science & Numerical Simulation,2016,35:17-24.
[4] 张群.金融市场复杂性与金融泡沫现象研究[D].广州:华南理工大学,2016.
[5] 赵利民,朱晓军.基于局部均值分解与样本熵的脑电信号特征提取与分类[J].计算机工程,2017,43(2):299-303.
[6] 孙曙光,于晗,杜太行,等.基于振动信号样本熵和相关向量机的万能式断路器分合闸故障诊断[J].电工技术学报,2017,32(7):20-30.
[7] 成娟,陈勋,彭虎.基于样本熵的肌电信号起始点检测研究[J].电子学报,2016,44(2):479-484.
[8] COSTA M,PENG C K,GOLDBERGER A L,et al.Multiscale entropy analysis of human gait dynamics[J].Physica A:Statistical Mechanics & Its Applications,2003,330(1-2):53-60.
[9] COSTA M,GOLDBERGER A L,PENG C K.Multiscale entropy analysis of complex physiologic time series[J].Physical Review Letters,2007,89(6):068102.
[10] COSTA M,HEALEY J A.Multiscale entropy analysis of complex heart rate dynamics:discrimination of age and heart failure effects[J].Computers in Cardiology,2003,30:705-708.
[11] 介丹.脑电信号的多尺度熵分析方法研究[D].太原:太原理工大学,2017.
[12] WU S D,WU C W,LEE K Y,et al.Modified multiscale entropy for short-term time series analysis[J].Physica A:Statistical Mechanics & Its Applications,2013,392(23):5865-5873.
[13] WU S D,WU C W,LIN S G,et al.Analysis of complex time series using refined composite multiscale entropy[J].Physics Letters A,2014,378(20):1369-1374.
[14] 仲兆满,胡云,李存华,等.微博中特定用户的相似用户发现方法[J].计算机学报,2016,39(4):765-779.
[15] 王晨旭,管晓宏,秦涛,等.微博消息传播中意见领袖影响力建模研究[J].软件学报,2015,26(6):1473-1485.
[16] HUMEAUHEURTIER A.The multiscale entropy algorithm and its variants:a review[J].Entropy,2016,17(5):3110-3123.
[2] 闫树熙.上证股市价格与成交量关系的实证研究[J].河南科学,2017,35(2):340-344.
[3] XU Mengjia,SHANG Pengjian,HUANG Jingjing.Modified generalized sample entropy and surrogate data analysis for stock markets[J].Communications in Nonlinear Science & Numerical Simulation,2016,35:17-24.
[4] 张群.金融市场复杂性与金融泡沫现象研究[D].广州:华南理工大学,2016.
[5] 赵利民,朱晓军.基于局部均值分解与样本熵的脑电信号特征提取与分类[J].计算机工程,2017,43(2):299-303.
[6] 孙曙光,于晗,杜太行,等.基于振动信号样本熵和相关向量机的万能式断路器分合闸故障诊断[J].电工技术学报,2017,32(7):20-30.
[7] 成娟,陈勋,彭虎.基于样本熵的肌电信号起始点检测研究[J].电子学报,2016,44(2):479-484.
[8] COSTA M,PENG C K,GOLDBERGER A L,et al.Multiscale entropy analysis of human gait dynamics[J].Physica A:Statistical Mechanics & Its Applications,2003,330(1-2):53-60.
[9] COSTA M,GOLDBERGER A L,PENG C K.Multiscale entropy analysis of complex physiologic time series[J].Physical Review Letters,2007,89(6):068102.
[10] COSTA M,HEALEY J A.Multiscale entropy analysis of complex heart rate dynamics:discrimination of age and heart failure effects[J].Computers in Cardiology,2003,30:705-708.
[11] 介丹.脑电信号的多尺度熵分析方法研究[D].太原:太原理工大学,2017.
[12] WU S D,WU C W,LEE K Y,et al.Modified multiscale entropy for short-term time series analysis[J].Physica A:Statistical Mechanics & Its Applications,2013,392(23):5865-5873.
[13] WU S D,WU C W,LIN S G,et al.Analysis of complex time series using refined composite multiscale entropy[J].Physics Letters A,2014,378(20):1369-1374.
[14] 仲兆满,胡云,李存华,等.微博中特定用户的相似用户发现方法[J].计算机学报,2016,39(4):765-779.
[15] 王晨旭,管晓宏,秦涛,等.微博消息传播中意见领袖影响力建模研究[J].软件学报,2015,26(6):1473-1485.
[16] HUMEAUHEURTIER A.The multiscale entropy algorithm and its variants:a review[J].Entropy,2016,17(5):3110-3123.