将生物信息学中的序列比对方法引入金融时间序列分析,可以捕获变量的大尺度特征,抑制噪声,并能从不同角度挖掘系统的隐含模式,且无需过度的前提假设.本文在已有序列比对方法的基础上,提出了两种用于金融序列比对的打分矩阵的构造方法,即相似度导向型矩阵和目的导向型矩阵,前者侧重于反映历史数据信息,可用于发现序列的对应模式,后者考虑对序列进行比对的目的,可用于提取序列的特征片段.应用该方法,本文对上证综指和深证成指的涨跌特征及其相关性进行了实证研究,得到了良好的研究效果,印证了将该方法引入金融领域分析的可行性和有效性.%Applying the sequence alignment method in bioinformatics to analyze financial time sequence can make it easy to capture large-scale features,remove noise and identify implicit pattern without having to rely too much on assumptions.This paper presents two methods of designing scoring matrix for financial sequence alignments,which are similarity-oriented matrix and purpose-oriented matrix.The former,focusing on the information of historical data,can be used to identify corresponding pattern.The latter,considering the pur pose of alignments,can be used to extract featured segments.In the empirical analysis,similarity-oriented and purpose-oriented matrixes are constructed to study the characteristics of ups and downs of Shanghai Composite Index,Shenzhen Component Index and the relationship between them.The satisfactory results verify the feasibility and effectiveness of the methodology in financial research.
展开▼
