• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2011, Vol. 33 ›› Issue (6): 150-153.

• 论文 • 上一篇    下一篇

基于PSO的属性选择方法

郑丽萍,姜华,李俊青   

  1. (聊城大学计算机学院,山东 聊城 252059)
  • 收稿日期:2010-08-29 修回日期:2010-12-05 出版日期:2011-06-25 发布日期:2011-06-25
  • 作者简介:郑丽萍(1979),女,山东聊城人,博士,讲师,研究方向为图像处理和人工智能。
  • 基金资助:

    山东省教育厅科研发展计划资助项目(J09LG29);聊城大学重点科研项目(X0810015)

A PSOBased Attribute Selection Method

ZHENG Liping,JIANG Hua,LI Junqing   

  1. (School of Computer Science,Liaocheng University,Liaocheng 252059,China)
  • Received:2010-08-29 Revised:2010-12-05 Online:2011-06-25 Published:2011-06-25

摘要:

为了减少实例对属性选择的影响,本文提出了基于PSO的属性选择方法。该方法主要利用PSO算法求实例群的最优熵值,获得相应的属性阈值,并利用阈值确定属性的优先级,最后按优先级进行选择。在实验中,通过确定本体中概念属性的优先级来验证所提算法的性能。实验结果表明,该方法减少了对实例的依赖,计算量也相对减少。

关键词: 微粒群算法, 属性优先级, 信息增益, 本体

Abstract:

In order to reduce the instance influence, a PSObased choosing method is proposed in this paper. This method mainly uses the  PSO algorithm to solve the optimal entropy of instances and obtain the corresponding attribute threshold value. According to the threshold, the attribute priority is determined. Finally, the attribute is chosen by priority. In our experiment, we specify the concept attribute priority in ontolgoy and verify the algorithm performance. The experimental results show that this algorithm reduces the dependency on instances and improves the accuracy. In addition, the computation quantity is reduced.

Key words: PSO;attribute priority;information gain;ontology