基于聚类的股票波动分析及其应用

Markowitz提出的“期望均值收益-收益方差“规则(M-V),模型要求选择差异性较大的资产进行组合,从而在给定收益率水平下,降低组合的风险。在M-V模型的基础上,采用了数据挖掘中聚类的方法,定义出一种衡量时间序列样本之间相似性程度的指标,这个指标反应了股票间波动行情趋势的异同。在此基础上对资产价格序列性进行聚类分析,与单纯M-V模型相比,在给定的收益率水平下降低了资产组合的风险。采用上证指数中若干股票进行实验验证表明,在给定的收益率下,采用基于密度的层次聚类方法的股票组合可以得到比随机组合更小的风险水平。

Fluctuation Analysis on Base of Clustering and Application

Markowitz proposed the rule of expected mean and the variance of return(M-V rule),which require as- sets with more difference to get a portfolio of lower risk given a certain return.In this paper a method of finding un- similar time series is presented to the fluctuant series on the base of data mining,and then choosing some proper as- sets to combine a portfolio.After obtaining the weight according to Markowitz theory,a portfolio with more return and lower risk with compare to Markowitz model are created.