股票数据流的相关性计算方法  

Computing Method of Correlation of Stock Data Streams

作　　者：彭宏[1] 刘洋[1] 邓维维[1] 郑启伦[1] 

机构地区：[1]华南理工大学计算机科学与工程学院,广州广东510640

出　　处：《华南理工大学学报（自然科学版）》2006年第1期86-89,共4页Journal of South China University of Technology(Natural Science Edition)

基　　金：广东省科技攻关资助项目(A10202001);广州市科技攻关资助项目(2004Z2-D0091)

摘　　要：相关性分析在股票投资、监测、预测中起着非常重要的作用.为定量计算各股票间的相关性,文中提出了一种股票数据流的相关性计算方法,它基于高效的单遍数据集扫描算法,能在有限的内存空间中计算出各股票间的相关性.与传统的计算方法相比,不管是在空间复杂度还是在时间复杂度上,所提出的方法都只需付出更小的代价.实验证明,这种相关性计算方法在同步股票报价中是有效的.Correlation analysis plays an important role in stock investment, detection and prediction. In order to quantitatively compute the correlation among different stocks, a method of computing the correlation among stock streams is proposed based on the efficient one-pass scanning algorithm. By using the proposed method the correlation between two stocks can be computed in limited memory. Moreover, the proposed method is of lower cost in both time complexity and space complexity than the traditional method, and is efficient in the synchronous quoting of stock price, which is verified by experiments.

关 键 词：股票 数据流 相关性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]