计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (29): 76-78.DOI: 10.3778/j.issn.1002-8331.2008.29.020

• 理论研究 • 上一篇    下一篇

基于非负矩阵分解方法的汉字基本部件识别

陈清华,陈六君,郑 涛,陈家伟   

  1. 北京师范大学 管理学院,北京 100875
  • 收稿日期:2007-11-26 修回日期:2008-01-16 出版日期:2008-10-11 发布日期:2008-10-11
  • 通讯作者: 陈清华

Base component discovery from Chinese characters by NMF and expanded NMF methods

CHEN Qing-hua,CHEN Liu-jun,ZHENG Tao,CHEN Jia-wei   

  1. School of Management,Beijing Normal University,Beijing 100875,China
  • Received:2007-11-26 Revised:2008-01-16 Online:2008-10-11 Published:2008-10-11
  • Contact: CHEN Qing-hua

摘要: 将NMF方法应用到汉字字形的处理中,成功地从一些汉字样本中抽取出构成这些汉字的基本部件。通过引入合适的惩罚因子,提出了一种扩展的NMF方法,对同样的汉字样本进行处理可以获得更好的结果,抽取出的基本部件就是构成这些汉字的偏旁部首。

关键词: 非负矩阵分解, 汉字, 基本部件

Abstract: The NMF method is applied to Chinese character and some base structures are discovered successfully.By introducing punishment factor into target function,a expended NMF method is proposed which is more suitable to factorize Chinese characters.

Key words: Nonnegative Matrix Factorization(NMF), Chinese characters, base component