• 论文 • 上一篇    下一篇

基于复杂网络的时间序列单变量波动幅度研究

安海岗,都沁军,张永礼   

  1. 石家庄经济学院,  石家庄 050031
  • 出版日期:2015-02-25 发布日期:2015-05-19

安海岗,都沁军,张永礼. 基于复杂网络的时间序列单变量波动幅度研究[J]. 系统科学与数学, 2015, 35(2): 158-169.

AN Haigang,DU Qinjun,ZHANG Yongli. RESEARCH ON THE FLUCTUATION RANGE OF SINGLE VARIABLE TIME SERIES BASED ON COMPLEX NETWORKS[J]. Journal of Systems Science and Mathematical Sciences, 2015, 35(2): 158-169.

RESEARCH ON THE FLUCTUATION RANGE OF SINGLE VARIABLE TIME SERIES BASED ON COMPLEX NETWORKS

AN Haigang ,DU Qinjun ,ZHANG Yongli   

  1. Shijiazhuang University of Economics, Shijiazhuang 05001
  • Online:2015-02-25 Published:2015-05-19

为研究时间序列单变量波动幅度演变规律,文章选择伦敦金下午收盘价格作为样本数据,借鉴统计物理学的方法进行研究. 利用粗粒化方法建立了价格波动幅度变化模态,运用复杂网络理论对时间序列单变量波动幅度模态的统计、变化规律和演化规律进行了分析.研究结果表明, 时间序列单变量波动幅度模态分布具有幂律性、群簇性和周期性,其波动幅度模态主要通过少数几种模态进行转换与演化.本研究方法不仅可以对不同类型时间序列单变量波动幅度进行研究, 同时可为多变量波动幅度及其联动波动规律研究提供思路.

In order to study the evolution law of the fluctuation range of single variable time series, this article chose closing price in the afternoon of London Gold from January 1, 2002 to February 28, 2014 as the sample data, in refer to the method of statistical physics research. Using Parkinson's definition about fluctuation range, it defines the calculation method of single variable fluctuation range, and uses coarse-grained method to transform price fluctuation range into symbolic sequences consisted of three characters $\{H,M,L\}$. The nodes of the fluctuation range complex network are 5-symbol strings in the form of data window linked in the network's topology by sliding sequence. The single variable fluctuation range complex network is composed of 149 models and the link edges between them. The former 32 nodes account for 79\% of the strength degree distribution. It does not show a good correlation between weighted clustering coefficient and point of strength. There are 10 small clusters that contain 8 points and above it in the network. The average shortest path length of network is 6.561, which means the price fluctuation range has periodicity. Some points that have higher degree have weaker betweenness centrality, however, some other points that have lower degree have an important role in the route of the network. In summary, the paper used the complex network theory to analyze the statistics, variation and evolution regularity of single variable time series fluctuation range mode. The result shows that the single variable time series fluctuation range has the characteristic of power-law,clustering and periodicity, and its fluctuation range models mainly apply a handful of models to transfer and evolve. The research results in this paper can not only provide methods for different types of single variable time series fluctuation range research, but also provide ideas for multiple variables fluctuation range and its linkage variation rule research.

MR(2010)主题分类: 

()
[1] 曹娟, 任凤丽. 耦合网络间的有限时聚类改进投影同步[J]. 系统科学与数学, 2021, 41(5): 1181-1190.
[2] 李盼盼, 董志良, 武天娇. 国际原油期货对中国新能源股指影响: 从多项式拟合到复杂网络[J]. 系统科学与数学, 2021, 41(5): 1355-1368.
[3] 董苏雅拉图. 证券市场中具有流动性人口特征的恐慌情绪传播模型[J]. 系统科学与数学, 2021, 41(10): 2919-2931.
[4] 李萍,倪志伟,朱旭辉,宋娟. 基于改进萤火虫算法的SVR空气污染物浓度预测模型[J]. 系统科学与数学, 2020, 40(6): 1020-1036.
[5] 周欢,刘嘉,马浩南. 改进的标签可重叠社区推荐模型[J]. 系统科学与数学, 2020, 40(11): 2058-2070.
[6] 石宇静,胡昌敏. 复杂网络的动态输出反馈容错同步控制[J]. 系统科学与数学, 2020, 40(10): 1701-1712.
[7] 肖峰,甘勤涛,黄欣. 具有多重权值的时滞复杂网络固定时间同步问题研究[J]. 系统科学与数学, 2020, 40(1): 15-28.
[8] 王甜,董志良,刘森,李盼盼. 原油价格时间序列自回归子模式传输特征分析[J]. 系统科学与数学, 2020, 40(1): 117-128.
[9] 赵尚威,周建红. 中国港口集装箱吞吐量预测:基于组合时间序列[J]. 系统科学与数学, 2018, 38(2): 210-219.
[10] 嵇敏,范玉涛,谢福鼎. 一种基于正交函数系的时间序列聚类方法[J]. 系统科学与数学, 2016, 36(1): 53-60.
[11] 郑晶,王应明,蓝以信. 考虑决策者心理行为的多时期应急决策方法[J]. 系统科学与数学, 2015, 35(5): 545-555.
[12] 陆刚. 农产品期货价格联动性实证研究------基于中美玉米期货日收盘价数据[J]. 系统科学与数学, 2015, 35(2): 181-192.
[13] 蓝以信,王应明. 考虑决策者时间偏好的多时期综合效率评价[J]. 系统科学与数学, 2014, 34(3): 257-272.
[14] 蓝以信,王应明. 考虑决策者时间偏好的多时期综合效率评价[J]. 系统科学与数学, 2014, 34(3): 257-272.
[15] 杨康,张仲义. 基于复杂网络理论的供应链网络风险传播机理研究[J]. 系统科学与数学, 2013, 33(10): 1224-1232.
阅读次数
全文


摘要