Title: An adaptive extreme learning machine algorithm and its application on face recognition

Authors: Jian Ni; Xinzheng Xu; Shifei Ding; Tongfeng Sun

Addresses: School of Computer Science and Technology, China University of Mining and Technology, No. 1, Daxue Road, Xuzhou, Jiangsu, China ' School of Computer Science and Technology, China University of Mining and Technology, No. 1, Daxue Road, Xuzhou, Jiangsu, China ' School of Computer Science and Technology, China University of Mining and Technology, No. 1, Daxue Road, Xuzhou, Jiangsu, China ' School of Computer Science and Technology, China University of Mining and Technology, No. 1, Daxue Road, Xuzhou, Jiangsu, China

Abstract: As a fast iterative neural networks learning algorithm, extreme learning machine (ELM) is a fast learning computation approach, and its whole learning process is completed through a mathematical transformation. However, the number of hidden layer neurons generally requires manual definition. In this paper, we propose an adaptive extreme learning machine (AP-ELM) algorithm which automatically determines the number of hidden layer neurons based on the AP clustering algorithm. In proposed algorithm, the samples are clustered through the AP clustering algorithm and then the number of cluster centres is used to determine the number of hidden layer neurons. Finally, the proposed adaptive ELM algorithm is used to face recognition. Experimental results verify that AP-ELM has good accuracy and reliability.

Keywords: extreme learning machine; ELM; clustering algorithms; classification; face recognition; biometrics; adaptive learning; neural networks.

DOI: 10.1504/IJCSM.2015.073601

International Journal of Computing Science and Mathematics, 2015 Vol.6 No.6, pp.611 - 619

Available online: 13 Dec 2015 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article