Optimised multikernels based extreme learning machine for face recognition Online publication date: Thu, 20-Dec-2018
by Bhawna Ahuja; Virendra P. Vishwakarma
International Journal of Applied Pattern Recognition (IJAPR), Vol. 5, No. 4, 2018
Abstract: Extreme learning machine (ELM) is a simple and efficient learning algorithm for regression and classification problems, because of its better generalisation performance. Though ELM is fast due to its mild optimisation constraints, the classification accuracy of basic ELM and its variants depend on the choice of kernel. Thorough investigations need to be done on conventional kernel-based classification algorithms to optimise the selection of kernels, which is essential for the good performance of a classifier. As different kernels have different feature extraction capability, a general framework to formulate an optimal kernel based on the concept of multiple kernels learning has been described in this paper. In the proposed approach, linearly combined base kernels have been used as the kernel matrix function which is non-static in nature. The values of coefficients used for combination can be tuned according to the application. Experiments on various face image datasets clearly reveal that the adoption of proposed framework can achieve better results than ELM and kernel ELM.
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Applied Pattern Recognition (IJAPR):
Login with your Inderscience username and password:
Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.
If you still need assistance, please email subs@inderscience.com