Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Recursive Training of Neural Networks for Classi cation
 

Summary: Recursive Training of Neural Networks for
Classi cation
Mayer Aladjem
Department of Electrical and Computer Engineering, Ben-Gurion University of the Negev,
P.O.B. 653, 84105 Beer-Sheva, Israel
List of Figures
1 AA network. It is trained to map input vectors into themselves in such a way that
yj
0 = yxj
0;w has overlapped class- conditional densities. . . . . . . . . . . . 4
2 Linear mapping y = wT x in a two- dimensional x-space. Parzen estimators
^pwT xj!1 and ^pwT xj!2 of the class-conditional densities along w. . . . . . . 6
3 Original data Xt into the x1;x2-plane. . . . . . . . . . . . . . . . . . . . . . 8
4 Transformed data Xt
0 after the RCS at 64 . . . . . . . . . . . . . . . . . . . . 8
5 Result of the method of Section 2. EPF w for various directions of w into
x1;x2-plane: original data path "|"; transformed data after successive RCS
at 64 path ":::" and 19 path "-.-.-". . . . . . . . . . . . . . . . . . . . . 8
6 Result of the method 2 . EPF w for various directions of w into x1;x2-plane:
original data path "|"; transformed data after successive RCS at 64 path

  

Source: Aladjem, Mayer - Department of Electrical and Computer Engineering, Ben-Gurion University

 

Collections: Computer Technologies and Information Sciences