Modified neural network activation function

Ibrahim, Adamu Abubakar and Chiroma, Haruna and Abdulkareem, Sameem and Ya’u Gital, Abdulsalam and Abdullahi Muaz, Sanah and Maitama, Jafaar and Isah, Muhammad Lamir and Herawan, Tutut (2014) Modified neural network activation function. In: 4th International Conference on Artificial Intelligence with Applications in Engineering and Technology (ICAIET 2014), 3 - 5th December, 2014, Kota Kinabalu, Sabah.

	PDF - Published Version Restricted to Repository staff only Download (354kB) | Request a copy
Preview	PDF Download (498kB) | Preview

Abstract

Neural Network is said to emulate the brain, though, its processing is not quite how the biological brain really works. The Neural Network has witnessed significant improvement since 1943 to date. However, modifications on the Neural Network mainly focus on the structure itself, not the activation function despite the critical role of activation function in the performance of the Neural Network. In this paper, we present the modification of Neural Network activation function to improve the performance of the Neural Network. The theoretical background of the modification, including mathematical proof is fully described in the paper. The modified activation function is code name as SigHyper. The performance of SigHyper was evaluated against state of the art activation function on the crude oil price dataset. Results suggested that the proposed SigHyper was found to improved accuracy of the Neural Network. Analysis of variance showed that the accuracy of the SigHyper is significant. It was established that the SigHyper require further improvement. The activation function proposed in this research has added to the activation functions already discussed in the literature. The study may motivate researchers to further modify activation functions, hence, improve the performance of the Neural Network.

Item Type:	Conference or Workshop Item (UNSPECIFIED)
Additional Information:	7132/39658 (ISBN: 978-1-4799-7910-3/14, DOI: 10.1109/ICAIET.2014.12 )
Uncontrolled Keywords:	— Artificial Neural Network; Activation Function; Training algorithm; Logsig; Hyperbolic
Subjects:	Q Science > Q Science (General) > Q300 Cybernetics > Q350 Information theory
Kulliyyahs/Centres/Divisions/Institutes (Can select more than one option. Press CONTROL button):	Kulliyyah of Information and Communication Technology Kulliyyah of Information and Communication Technology
Depositing User:	Dr Adamu Abubakar
Date Deposited:	19 Dec 2014 16:30
Last Modified:	23 May 2018 14:35
URI:	http://irep.iium.edu.my/id/eprint/39658

Actions (login required)

View Item