计算机工程与应用 ›› 2012, Vol. 48 ›› Issue (34): 40-44.

• 研究、探讨 • 上一篇    下一篇

FKA算法迭代收敛性分析

刘忠宝1,2,赵文娟3   

  1. 1.中北大学 电子与计算机科学技术学院,太原 030051
    2.江南大学 物联网工程学院,江苏 无锡 214122
    3.山西大学 商务学院 信息学院,太原 030031
  • 出版日期:2012-12-01 发布日期:2012-11-30

Iterative convergence analysis on algorithm FKA

LIU Zhongbao1,2, ZHAO Wenjuan3   

  1. 1.School of Electronics and Computer Science Technology, North University of China, Taiyuan 030051, China
    2.School of IoT Engineering, Jiangnan Univerisity, Wuxi, Jiangsu 214122, China
    3.Information School, Business College of Shanxi University, Taiyuan 030031, China
  • Online:2012-12-01 Published:2012-11-30

摘要: 《核选择和非线性特征提取的双线性分析》一文提出了一种新颖的核Fisher准则FKC, 并用迭代分析算法FKA求得最优解,但其迭代收敛性缺乏理论上的证明。从理论上对FKA算法的迭代收敛性进行了分析和探讨,并运用Radermacher复杂性分析法进行证明。

关键词: 迭代收敛性分析, 线性判别分析, Radermacher复杂性分析, 核Fisher准则

Abstract: In the paper “bilinear analysis for kernel selection and nonlinear feature extraction” (IEEE Trans on NN, 2007, 18(5)), the authors presented a unified criterion, Fisher and Kernel Criterion(FKC), for feature extraction and recognition and used an iterative procedure to optimize the new criterion. But there is still no theoretical discussion concerning the convergence issue of such an iterative procedure. An iterative convergence analysis of Fisher and Kernel analysis algorithm(FKA) is povided using the concept of Radermacher complexity.

Key words: iterative convergence analysis, Linear Discriminant Analysis, Radermacher complexity, Fisher and Kernel Criterion(FKC)