Fitting of scaling curve and financial time series clustering

doi:10.11772/j.issn.1001-9081.2014.11.3344

Abstract

Abstract:

In order to take the multi-fractal properties of financial time series into consideration, a clustering method based on measuring similarities among CSI 300 index stocks through scaling curve was proposed. The algorithm firstly fitted scaling curve at different autocorrelation order through Multi-Scale Detrend Fluctuation Analysis (MSDFA). Then it abstracted the distribution or shape features of scaling curve for the construction of pattern vector. Clustering was implemented by weighted K-means algorithm and the optimal number of categories was determined by Davis-Bouldin Index (DBI). The result shows the clustering based on scaling curve can discover industry aggregation and strong linkage of different plates within the stock market. The portfolio built from different clusters can reduce the risk greatly and the proposed method outperforms clustering method based on linear trend features of original series.

摘要：

针对金融时间序列具有的多重分形特征,提出基于标度曲线测度沪深300指标股之间的相似性并实现聚类。该方法首先使用多标度退势波动分析(MSDFA)拟合不同自相关阶数下收益率序列的标度曲线,然后抽取其分布或形态特征构造模式向量。聚类通过含权K-means算法实现,最优类别数根据分类适确性指标(DBI)确定。结果显示,基于标度曲线的聚类能够揭示出股市的行业聚集性和板块间的关联性,在此基础上构造的投资组合可以显著降低风险,并且效果优于基于原始序列线性趋势特征的聚类。

CLC Number:

TP391.4

YUAN Ming. Fitting of scaling curve and financial time series clustering[J]. Journal of Computer Applications, 2014, 34(11): 3344-3347.

袁铭. 标度曲线拟合与金融时间序列聚类[J]. 计算机应用, 2014, 34(11): 3344-3347.

References

[1]ZHANG Z, SUN Y, ZHANG S, et al. Clustering analysis of electric load series using clustering algorithm of multi-hierarchy and detailed decomposition based on data mining[J]. Power System Technology,2006,30(2):51-56.(张智晟,孙雅明,张世英,等. 基于数据挖掘多层次细节分解的负荷序列聚类分析[J]. 电网技术,2006,30(2):51-56.)
[2]WANG L,TENG S. Application of clustering and time-based sequence analysis in intrusion detection[J]. Journal of Computer Applications,2010,30(3):699-701,714.(王令剑,滕少华. 聚类和时间序列分析在入侵检测中的应用[J]. 计算机应用,2010,30(3):699-701,714.)
[3]CHEN M, CHEN M, WU W, et al. Dynamic casual relationship in different industry indices: China's evidence[J]. Systems Engineering — Theory and Practice, 2009,29(6):19-31.(陈暮紫,陈敏,吴武清,等. 中国A股市场行业板块间领滞关系的动态变化实证研究[J]. 系统工程理论与实践,2009,29(6):19-31.)
[4]LI Q, ZHANG T, DING H. Herd behavior of the securities investment funds in China-empirical analysis based on periodic law and plate effect[J]. Journal of Central University of Finance and Economics,2013(6):37-43.(李奇泽,张铁刚,丁焕强. 中国证券投资基金羊群行为——基于周期规律与板块效应的实证分析[J]. 中央财经大学学报,2013(6):37-43.)
[5]CAI S, XIA Z, ZHANG W. Semi-supervised spectral clustering of time-series similarity[J]. Computer Engineering and Applications, 2011,47(31):116-118.(蔡世玉,夏战国,张文涛. 时间序列相似性半监督谱聚类[J]. 计算机工程与应用,2011,47(31):116-118.)
[6]KALPAKIS K, GADA D, PUTTAGUNTA V. Distance measures for effective clustering of ARIMA time series[C]// Proceedings of the 2001 IEEE International Conference on Data Mining. Piscataway: IEEE Press, 2001:273-280.
[7]ZHANG H, HO T, LIN M. A non-parametric wavelet feature extractor for time-series classification[C]// KAKDD 2004: Proceedings of the 8th Pacific-Asia Conference, LNCS 3056. Berlin: Springer-Verlag, 2004:595-603.
[8]VAITHYANATHAN S, DOM B. Model-based hierarchical clustering[EB/OL]. [2010-10-10]. http://arxiv.org/ftp/arxiv/papers/1301/1301.3899.pdf.
[9]KANTELHARDT J W, ZSCHIEGNER S A, KOSCIELNY-BUNDE E, et al. Multifractal detrended fluctuation analysis of nonstationary time series[J]. Physica A: Statistical Mechanics and its Applications, 2002,316(1/2/3/4):87-114.
[10]ZHANG L, LIU C. Multifractal analysis of Japan and China stock markets in different economy Periods[J]. Systems Engineering — Theory and Practice,2013,33(2):317-328.(张林,刘春燕. 日中两国不同经济时期股市的多重分形分析[J]. 系统工程理论与实践,2013,33(2):317-328.)
[11]HUANG J, CHENG H, GUO Y, et al. Research on multifractal features of the relation between price and volume in China metal futures market: based on MF-DCCA approach[J]. Management Review,2013,25(4):77-85.(黄健柏,程慧,郭尧琦,等. 金属期货量价关系的多重分形特征研究——基于MF-DCCA方法[J]. 管理评论,2013,25(4):77-85.)
[12]NI Z, HU T, WU X, et al. A novel machine learning approach based on fractal theory: Fractal learning[J]. Journal of University of Science and Technology of China,2013,43(4):265-270.(倪志伟,胡汤磊,吴晓璇,等. 基于分形理论的一种新的机器学习方法:分形学习[J]. 中国科学技术大学学报,2013,43(4):265-270.)
[13]HUANG C, WU Q, WU Z, et al. Clustering financial time series based on multi-fractal features of variance volatility[J]. System Engineering, 2006,24(6):100-103.(黄超,吴清烈,武忠,等. 基于方差波动多重分形特征的金融时间序列聚类[J].系统工程,2006,24(6):100-103.)
[14]ZHONG W, GAO Q, CHENG Y. The clustering of financial time serials based on wavelet and multi-fractal[J]. Systems Engineering,2009,27(3):58-61.(钟维年,高清维,陈燕玲. 基于小波和多重分形的金融时间序列聚类[J]. 系统工程,2009,27(3):58-61.)
[15]YIN Y, SHAN P. Modified DFA and DCCA approach for quantifying the multiscale correlation structure of financial markets[J]. Physica A: Statistical Mechanics and its Applications,2013,392(24): 6442-6457.