TOWARDS GLOBAL OPTIMIZATION OF NEURAL NETWORK: A COMPARATIVE ANALYSIS USING GENETIC AND WHALE OPTIMIZATION ALGORITHMS

The intelligence and adaptability features of the neural network has made it a technique that is widely used to solve problems in diverse areas such as; detection, monitoring, prediction, diagnostics, data mining, classification, recognition, robotics, biomedicine, etc. However, determination of the...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Journal of biomedical engineering and medical imaging Jg. 8; H. 6; S. 89 - 101
Hauptverfasser: Igodan, E. C., Ukaoha, K. C., Oliomogbe, S. O. P.
Format: Journal Article
Sprache:Englisch
Veröffentlicht: 14.12.2021
ISSN:2055-1266, 2055-1266
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The intelligence and adaptability features of the neural network has made it a technique that is widely used to solve problems in diverse areas such as; detection, monitoring, prediction, diagnostics, data mining, classification, recognition, robotics, biomedicine, etc. However, determination of the optimal number of hidden layers of neural network and other parameters are still a difficult task. Usually, these parameters are decided by trial-and-error which increases the computational complexity and it is human dependent in obtaining the optimal model and parameters alike for any particular task. Optimization has received enormous attention in recent years, primarily because of the rapid progress in computer technology, including the development and availability of user-friendly software, high-speed and parallel processors, and artificial neural networks. This research work is to propose a neuro-evolutionary model using the computational intelligence techniques by combining ANN, GA and WOA for binary classification problems. The proposed optimized ANN-GA and WOA models is to circumvent the problem that is characterized in the trade-off between smoothness and accuracies in selecting the models and optimal parameters of neural network.
ISSN:2055-1266
2055-1266
DOI:10.14738/jbemi.86.11004