分形视角下股票市场的聚类分析

Cluster Analysis of Stock Market from Fractal Perspective

针对金融时间序列具有的多重分形特征,从分形的角度出发,对股票市场进行分析.基于多重分形指数分布和多重分形谱的两个特殊值,以传统欧式距离函数为测度函数,使用层次聚类法,对股票市场的金融时间序列进行聚类分析.研究结果表明,股票市场存在多重分形特征,且聚类结果显示,将股票分为绩优股、垃圾股、潜力股3大类较为合理.

Aiming at the multifractal characteristics of financial time series,this paper analyzes the stock market from the fractal perspective.Based on the multifractal index distribution and the two special values of multifractal spectrum,the traditional Euclidean distance function is used as the measure function,and the hierarchical clustering method is used to cluster the financial time series of the stock market.The results show that there are multifractal features in China’s stock market,and the clustering results show that it is reasonable to classify the stock into three categories:bule-chip stocks,junk stocks and potential stocks.