Computer Engineering and Applications ›› 2012, Vol. 48 ›› Issue (31): 166-170.

Previous Articles     Next Articles

Pure parameterized extension of pattern trees induction

YI Yu1, LIAO Niandong1, HUANG Jingquan2   

  1. 1.School of Computer Science, Changsha University of Science and Technology, Changsha 410114, China
    2.Academic Administration’s Office, Helong Sports School, Changsha 410151, China
  • Online:2012-11-01 Published:2012-10-30

模式树推导方法的纯参数化扩展

易  禹1,廖年冬1,黄景全2   

  1. 1.长沙理工大学 计算机与通信工程学院,长沙 410114
    2.长沙市贺龙体育运动学校 教务处,长沙 410151

Abstract: A new linear way to construct a parameterized fuzzy AND/OR norm is introduced by analyzing the pattern tree induction method, which makes pattern tree a purely parameterized model. The new model has high accuracy and good readability and interpretability on the one hand, and enjoys the great flexibility on the other hand. Correspondingly, a general optimization technique is employed to tune pattern tree, the experimental results show that the accuracy of new pattern tree is increased effectively.

Key words: pattern tree, fuzzy classification, fuzzy norm, parameterization

摘要: 通过分析模式树推导方法的特点,提出一种新的线性方法来构造参数化的模糊与/或算子,使得模式树成为一个完全由参数化模糊算子构成的模型。该方法既保持了原有模型可读性强、精度高和可解释性强的优点,又有效地提高了该模型的灵活性;同时建立了相应的参数优化机制。实验结果表明,新的模式树在分类精度上比原有模型有了较显著的提高。

关键词: 模式树, 模糊分类, 模糊算子, 参数化