Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (14): 207-209.DOI: 10.3778/j.issn.1002-8331.2009.14.064

• 工程与应用 • Previous Articles     Next Articles

Application of PCA based on kernel function in analysis of QAR data

FENG Xing-jie,FENG Xiao-rong,WANG Yan-hua   

  1. School of Computer Science & Technology,Civil Aviation University of China,Tianjin 300300,China
  • Received:2008-03-17 Revised:2008-06-10 Online:2009-05-11 Published:2009-05-11
  • Contact: FENG Xing-jie

基于核函数的PCA在QAR数据分析中的应用

冯兴杰,冯小荣,王艳华   

  1. 中国民航大学 计算机科学与技术学院,天津 300300
  • 通讯作者: 冯兴杰

Abstract: This paper analyzes the drawbacks of general Principal Component Analysis(PCA) firstly,and discusses the Kernel Principal Component Analysis(KPCA) and its drawbacks of high time complexity secondly.Then proposes the kernel function covariance matrix of principal component analysis in the end.Compared to KPCA,the method is fast descending dimension speed.The results show that the proposed method used for QAR data has a good effect of dimension reduction and high rate of correct classification.

Key words: Principal Component Analysis(PCA), kernel function, Kernel Principal Component Analysis(KPCA), covariance matrix

摘要: 分析了传统的主成分分析方法的不足,论述了KPCA方法及其时间复杂度高的缺陷。在此基础上,提出基于核函数构造的协方差矩阵的主成分分析,相比 KPCA,该方法具有快的降维速度。实验结果显示:把该方法用于QAR数据具有良好的降维效果和高分类正确率。

关键词: 主成分分析, 核函数, 核主成分分析, 协方差矩阵