UTHM Institutional Repository

Improving the accuracy of gradient descent back propagation algorithm (GDAM) on classification problems

Rehman, M. Z. and Mohd Nawi, Nazri Improving the accuracy of gradient descent back propagation algorithm (GDAM) on classification problems. International Journal on New Computer Architectures and Their Applications (IJNCAA), 1 (3). pp. 861-870. ISSN 2220-9085

Full text not available from this repository.

Abstract

The traditional Back-propagation Neural Network (BPNN) Algorithm is widely used in solving many real time problems in world. But BPNN possesses a problems in world. But BPNN possesses a problem of slow convergence and convergence to local minima. Previously, several modifications are suggested to improve the convergence rate the Gradient Descent Back-propagation algorithm such as careful selection of initial weights and biases, learning rate, momentum, network topology, activation function and 'gain' value in the activation function. This research proposed an algorithm for improving the current working performance of Back-propagation algorithm by adaptively changing the momentum value and at the same time keeping the 'gain' parameter fixed for all nodes in the neural network. The performance of the proposed method known as 'Gradient Descent Method with Adaptive Momentum (GDAM)' is compared with the performances of 'Gradient Descent Method with Adaptive Gain (GDAM-AG)' and 'Gradient Descent with Simple Momentum (DGM)'. The efficiency of the purposed method is demonstrated by simulations on five classification problems. results shown that GDAM can be used as an alternative approach for BPNN because it demonstrate better accuracy ratio on the chosen classification problems.

Item Type: Article
Uncontrolled Keywords: gradient descent; neural network; adaptive momentum; adaptive gain; back-propagation
Subjects: Q Science > QA Mathematics > QA75 Calculating machines > QA75.5 Electronic computers. Computer science
Divisions: Faculty of Computer Science and Information Technology > Department of Software Engineering
Depositing User: Normajihan Abd. Rahman
Date Deposited: 07 Feb 2013 05:53
Last Modified: 22 Jan 2015 03:52
URI: http://eprints.uthm.edu.my/id/eprint/2965
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item