Journal of Systems Engineering and Electronics ›› 2009, Vol. 31 ›› Issue (3): 583-587.

• 系统工程 • 上一篇    下一篇

一种大规模数据库的组合优化决策树算法

赵静娴1,2, 倪春鹏1, 詹原瑞1, 杜子平2   

  1. 1. 天津大学管理学院, 天津, 300072;
    2. 天津科技大学经管学院, 天津, 300222
  • 收稿日期:2007-10-22 修回日期:2007-12-20 出版日期:2009-03-20 发布日期:2010-01-03
  • 作者简介:赵静娴(1981- ),女,博士研究生,主要研究方向为金融工程与数据挖掘.E-mail:nzjx2005@163.com
  • 基金资助:
    国家自然科学基金(70573076;70671074);天津科技大学科研基金(20080303)资助课题

Combined optimization decision tree algorithm suitable for large scale data-base

ZHAO Jing-xian1,2, NI Chun-peng1, ZHAN Yuan-rui1, DU Zi-ping2   

  1. 1. School of Management, Tianjin Univ., Tianjin 300072, China;
    2. School of Economics and Management, Tianjin Univ. of Science & Technology, Tianjin 300222, China
  • Received:2007-10-22 Revised:2007-12-20 Online:2009-03-20 Published:2010-01-03

摘要: 提出了一种适合于大规模高维数据库的组合优化决策树算法。相比于传统的类似算法,该算法从数据的离散化,降维,属性选择三方面进行改进,对决策树建立过程中不适应大规模高维数据库的主要环节进行了优化,有效解决了处理大规模高维数据库问题的效率和精度之间的矛盾。仿真试验表明,该算法在大大减少了计算代价的同时提高了决策树的分类精度。

Abstract: A combined optimization decision tree algorithm suitable for a large scale and high dimension data-base is presented.Compared with the traditional similar algorithms,the algorithm makes improvements from three aspects: discretization,reducing dimension and attribute selection.It also optimizes the main processes,so that it is suitable for large scale and high dimension data-base and effectively solves the conflict between efficiency and predictive precision.Experiments show that the proposed method raises the predictive precision of decision trees while it greatly reduces the computational cost.

中图分类号: