时序数据涨落分形维数测度研究——以股票市场为例  

Research on Measuring Fluctuation Fractal Dimension of Time Series Data——Taking Stock Markets as Examples

作　　者：鄢章华[1] 吕衍轶 刘蕾 洪璇娇 Yan Zhanghua;Lyu Yanyi;Liu Lei;Hong Xuanjiao(School of Management,Harbin Commerce University,Harbin,China;School of Economics and Management,Taizhou University,Taizhou,China)

机构地区：[1]哈尔滨商业大学管理学院,黑龙江哈尔滨 [2]泰州学院经济与管理学院,江苏泰州

出　　处：《科学技术创新》2024年第22期93-96,共4页Scientific and Technological Innovation

基　　金：黑龙江省哲学社会科学研究规划项目(22GLB119);哈尔滨商业大学镜湖学者支持计划(JHQNRC03);泰州学院引进人才科研启动项目(TZXY2019QDJJ005)。

摘　　要：为了更好分析时序数据的规律与特征,文章基于分形理论设计时间序列数据的涨落分形维数的测度方法,并以中美股票交易市场的股价数据为例,从测度单一股票的分形维数到分析整个市场的分形维数分布特征,说明算法的应用方式。算法应用研究发现,中国股票市场上的股票涨落的分形维数数据服从两种正态分布的混合分布,符合“政府+市场”的中国特色,且股价涨落的分形维数还反映了中国的两个股票交易市场的差异化定位与功能,从侧面验证了算法的有效性。In order to better analyze the rule and characteristics of time series data,this paper designs a method to measure the fluctuation fractal dimension of time series data based on fractal theory,from calculating the fractal dimension of a single stock to analyzing the distribution characteristics of the fractal dimension of the whole market,the application of the algorithm is illustrated.The study of algorithm application shows that the fractal dimension data of stock fluctuation in China's stock market obey the mixed distribution of two normal distributions,which accords with the Chinese characteristics of"government+market",the fractal dimension of stock price fluctuation also reflects the differential positioning and function of China's two stock exchange markets,which verifies the effectiveness of the algorithm.

关 键 词：分形维数 股价涨落 中美对比 分布 

分 类 号：F830.91[经济管理—金融学]