Enhancing the Performance of the BackPropagation for Deep Neural Network
Journal Title: INTERNATIONAL JOURNAL OF COMPUTERS & TECHNOLOGY - Year 2014, Vol 13, Issue 12
Abstract
The standard Backpropagation Neural Network (BPNN) Algorithm is widely used in solving many real problems in world. But the backpropagation suffers from different difficulties such as the slow convergence and convergence to local minima. Many modifications have been proposed to improve the performance of the algorithm such as careful selection of initial weights and biases, learning rate, momentum, network topology and activation function. This paper will illustrate a new additional version of the Backpropagation algorithm. In fact, the new modification has been done on the error signal function by using deep neural networks with more than one hidden layers. Experiments have been made to compare and evaluate the convergence behavior of these training algorithms with two training problems: XOR, and the Iris plant classification. The results showed that the proposed algorithm has improved the classical Bp in terms of its efficiency.
Authors and Affiliations
Ola Mohammed Surakhi, Walid A. Salameh
Keywords
Related Articles
- EP ID EP650625
- DOI 10.24297/ijct.v13i12.5279
- Views 80
- Downloads 0
Information Hiding Using Mosaic Images with Random Image Key Generation
Image watermarking is one of the efficient technique to send the secrete information from sender to receiver. The concept of information hiding using mosaic images is also an efficient technique of keeping the informatio...
DESIGN OF CMOS LOW DROP-OUT REGULATORS: A COMPARATIVE STUDY
The advancement in battery operated portable devices, noise sensitive devices and other devices, which need high precision supply voltages has fuelled the growth of Low Drop-Out Regulators. Low Drop-Out Regulators showed...
NondestructiveApproachforDetermination of Steel MechanicalProperties
It was proposed the design of an artificial neural network (ANN) to estimate the yield strength in the welding zone of HSLA experimental steels. The input parameters of the ANN were the chemical composition and hardness....
CONCEPTUAL THREE PHASE ITERATIVE MODEL OF KDD
KDD process includes how data is stored and accessed, how andwhat algorithms can apply to large amount of data efficiently,how results can be interpreted and visualized. KDD is theprocess of identifying valid, interestin...
Fuzzy in the Real World
The Fuzzy Logic tool was introduced in 1965, by LotfiZadeh, and is a mathematical tool for dealing with uncertainty. It offers to a soft computing partnership the important concept of computing with words. It provides a...