Journal of South China University of Technology (Natural Science Edition) ›› 2006, Vol. 34 ›› Issue (1): 86-89.

• Computer Science & Technology • Previous Articles     Next Articles

Computing Method of Correlation of Stock Data Streams

Peng Hong  Liu Yong  Deng Wei-wei  Zheng Qi-lun   

  1. College of Computer Science and Engineering,South China Univ.of Tech.,Guangzhou 510640,Guangdong,China
  • Received:2005-04-05 Online:2006-01-25 Published:2006-01-25
  • Contact: 彭宏(1956-)，教授，博士生导师，主要从事智能网络技术、智能商务与数据挖掘、人工智能应用技术方面的研究 E-mail:mahpeng@scut.edu.cn
  • About author:彭宏(1956-)，教授，博士生导师，主要从事智能网络技术、智能商务与数据挖掘、人工智能应用技术方面的研究
  • Supported by:

    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)

PDF

1041

Abstract:

Correlation analysis plays an important role in stock investment,detection and prediction.In order to quantitatively compute the corelation among different stocks,a method of computing the corelation among stock streams is proposed based orl the efficient one-pass scanning algorithm.By using the proposed method the corela-tion 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.

Key words: stock, data stream, corelation