Computer Engineering and Applications ›› 2010, Vol. 46 ›› Issue (20): 89-92.DOI: 10.3778/j.issn.1002-8331.2010.20.025

• 数据库、信号与信息处理 • Previous Articles     Next Articles

Seperability order based classification feature extraction method for functional data

WAN Jun1,2,GE Bing-feng1,CHEN Ying-wu1,BAI Zhi-dong2   

  1. 1.College of Information System and Management,National University of Defense Technology,Changsha 410073,China
    2.Department of Statistic and Probability Application,Faculty of Science,National Uinversity of Singapore,117546,Singapore
  • Received:2010-03-10 Revised:2010-04-27 Online:2010-07-11 Published:2010-07-11
  • Contact: WAN Jun

基于可分度序的函数化数据分类特征提取方法

万 君1,2,葛冰峰1,陈英武1,白志东2   

  1. 1.国防科技大学 信息系统与管理学院,长沙 410073
  • 通讯作者: 万 君

Abstract: The most different character of Functional Data(FD) is the high dimension and high correlation.So,it is a key problem to extract features from FD while keeping its global characters,which relates to the classification efficiency and precision to heavens.A novel method using K-L seperability order to extract classification features and reduce the demision is proposed which using information theory idea for reference.This method is an advanced approach of the popular wavelet based shrinkage method for functional data reduction.It is proved by theory analysis and experiment test that this method has advantages in improving classification efficiency,precision and robustness.

Key words: functional data, feature extraction, wavelet, classification, K-L distance, seperability

摘要: 由于函数化数据的高维、高相关性特点,如何在保持其整体特性的前提下提取函数化数据的分类特征,是关系到能否有效提高分类效率和精度的关键问题。改进了当前常用的基于小波阀值法的函数化逐步降维方法,针对分类问题,借鉴信息论的思想,采用K-L可分度排序法构建了新的分类特征提取与降维规则。理论分析和实验表明,该方法能有效提取分类特征,提高分类效率、分类精度和分类稳健性。

关键词: 函数化数据, 特征提取, 小波, 分类, 可分度

CLC Number: