Analysis of feature extraction by inverse mapping and Alopex algorithm

Author

Nagashino, Hirofumi ; Yamamoto, Hidenori ; Pandya, Abhijit S. ; Kinouchi, Yohsuke

Author_Institution

Dept. of Electr. & Electron. Eng., Tokushima Univ., Japan

Volume

4

fYear

1994

fDate

27 Jun-2 Jul 1994

Firstpage

2407

Abstract

This paper describes two methods of analysis of feature extraction characteristics of pattern recognition neural networks and their application to a three-layer feedforward network that learns by backpropagation. One method is inverse mapping and the other is calculation by Alopex algorithm, which is an iterative and stochastic processing to minimize or maximize a cost function. By these methods the receptive fields of the units in the hidden and output layers are obtained. Effectiveness of these methods are demonstrated

Keywords

backpropagation; feature extraction; feedforward neural nets; iterative methods; minimax techniques; stochastic processes; Alopex algorithm; backpropagation; cost function; feature extraction; inverse mapping; iterative method; neural networks; pattern recognition; stochastic processing; three-layer feedforward network; Backpropagation algorithms; Cost function; Feature extraction; Feedforward neural networks; Iterative algorithms; Iterative methods; Neural networks; Pattern analysis; Pattern recognition; Stochastic processes;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 1994. IEEE World Congress on Computational Intelligence., 1994 IEEE International Conference on

Conference_Location

Orlando, FL

Print_ISBN

0-7803-1901-X

Type

conf

DOI

10.1109/ICNN.1994.374596

Filename

374596