Adaptive Regularizer for Recursive Neural Network Training Algorithms

Author

Asirvadam, Vijanth S.

Author_Institution

Dept. of Electr. & Electr. Eng., Univ. Teknol. Petronas, Tronoh

fYear

2008

fDate

16-18 July 2008

Firstpage

89

Lastpage

94

Abstract

Adaptive Marquardt parameter correction techniques are tested for recursive Levenberg-Marquardt (RLM) and proposed novel application on decomposed recursive Levenberg Marquardt (DRLM) algorithms. The adaptive Marquardt correction is based on recursive moving-window residual. Experiment results show superior convergence using decomposed approach and a slight improvement in performance by adopting the adaptive Marquardt correction ona fixed size multilayer perceptrons (MLP) network.

Keywords

learning (artificial intelligence); multilayer perceptrons; adaptive Marquardt parameter correction techniques; adaptive regularizer; decomposed recursive Levenberg Marquardt algorithms; multilayer perceptron network; recursive moving-window residual; recursive neural network training algorithms; Chaos; Computer networks; Conferences; Convergence; Cost function; Multilayer perceptrons; Neural networks; Neurons; Recursive estimation; Testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Computational Science and Engineering Workshops, 2008. CSEWORKSHOPS '08. 11th IEEE International Conference on

Conference_Location

San Paulo

Print_ISBN

978-0-7695-3257-8

Type

conf

DOI

10.1109/CSEW.2008.55

Filename

4625045