Evaluating the process of a genetic algorithm to improve the back-propagation network: A Monte Carlo study

ASIA unversity > 資訊學院 > 行動商務與多媒體應用學系 > 期刊論文 > Item 310904400/8721

Please use this identifier to cite or link to this item: http://asiair.asia.edu.tw/ir/handle/310904400/8721

Title:	Evaluating the process of a genetic algorithm to improve the back-propagation network: A Monte Carlo study
Authors:	Huang, Chien-Yu;Chen, Long-Hui;Chen, Yueh-Li;Chang, Fengming M.
Contributors:	Department of Information Science and Applications
Keywords:	Algorithms;Assembly;Backpropagation;Genetic algorithms;Image classification;Learning systems;Linearization;Monte Carlo methods;Topology;Algorithm operations;Back-propagation network;Back-propagation networks;Gradient searches;Monte Carlo studies;Neural network trainings;Nonlinear optimization problems;Nonlinear problems;Overfitting;Prediction errors;Solution spaces;Test functions;Training errors
Date:	2009-03
Issue Date:	2010-04-07 13:34:10 (UTC+0)
Publisher:	Asia University
Abstract:	Many studies have mapped a bit-string genotype using a genetic algorithm to represent network architectures to improve performance of back-propagation networks (BPN). But the limitations of gradient search techniques applied to complex nonlinear optimization problems have often resulted in inconsistent and unpredictable performance. This study focuses on how to collect and re-evaluate the weight matrices of a BPN while the genetic algorithm operations are processing in each generation to optimize the weight matrices. In this way, overfitting, a drawback of BPNs that usually occurs during the later stage of neural network training with descending training error and ascending prediction error, can also be avoided. This study extends the parameters and topology of the neural network to enhance the feasibility of the solution space for complex nonlinear problems. The value of the proposed model is compared with previous studies using a Monte Carlo study on in-sample, interpolation, and extrapolation data for six test functions. © 2007 Elsevier Ltd. All rights reserved.
Relation:	Expert Systems with Applications 36(2) PART 1:1459-1465
Appears in Collections:	[行動商務與多媒體應用學系] 期刊論文

Files in This Item:

File	Description	Size	Format
		0Kb	Unknown	317	View/Open
430.doc		37Kb	Microsoft Word	227	View/Open

Loading...