作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2008, Vol. 34 ›› Issue (3): 225-227. doi: 10.3969/j.issn.1000-3428.2008.03.080

• 人工智能及识别技术 • 上一篇    下一篇

基于代数特征与粗糙集相结合的特征提取方法

邵 俊1,吴小俊1,2,王士同2,杨静宇3   

  1. (1. 江苏科技大学电子与信息学院,镇江 212003;2. 江南大学信息工程学院,无锡 214122;3. 南京理工大学信息学院,南京 210094)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2008-02-05 发布日期:2008-02-05

基于代数特征与粗糙集相结合的特征提取方法

SHAO Jun1, WU Xiao-jun1,2, WANG Shi-tong2, YANG Jing-yu3   

  1. (1. School of Electronics and Information, Jiangsu University of Science and Technology, Zhenjiang 212003; 2. School of Information Engineering, Southern Yangtse University, Wuxi 214122; 3. School of Information, Nanjing University of Science and Technology, Nanjing 210094)
  • Received:1900-01-01 Revised:1900-01-01 Online:2008-02-05 Published:2008-02-05

摘要: 对特征抽取方法进行了研究,提出一种新的特征抽取方法,克服了Roman W等提出的特征抽取方法中缺乏鉴别信息的缺点。通过对高维的人脸数据用PCA和LDA降维,利用粗糙集理论中的属性约简算法进行进一步的维数压缩。实验结果表明,该方法具有良好的性能。

关键词: 属性约简, 属性重要度, 近似约简误差

Abstract: A study is made on feature extraction and a new feature extraction method is suggested. The method overcomes the shortcomings of the feature extraction method which was proposed by Roman W recently, which has no discriminant information. Face feature is obtained by the combination of PCA and LDA, the dimensionality is further reduced by using attribute reduction algorithm in the theory of rough set. Experimental results indicate that the proposed method has good performance.

Key words: attributes reduction, attribute significance, error of reduct approximation

中图分类号: