Flash News
Welcome to IARC

Publisher Login

Latest News
Welcome to IARC- JCR Report

Submit your Journal to get IARC-JCRR Indexing and Impact Factor

Impact Factor calculated by IARC on the basis of Journal Citation Reference (JCR) Report.


Contact: iarcdpi@gmail.com


Efficient Multi Kernel Implementation optimised with Weight Updation Algorithm
Kanchana .J , Gladston Raj .S

Published in: Journal for Advance Research in Applied Sciences
Volume- 4, Issue-7, pp.201-211, Dec 2017
DPI :-> 16.10089.JARAS.2017.V4I7.201211.2239

Support Vector Machine (SVM) has a vital role in machine learning applications in this emerging era of knowledge mining. The drawbacks in selecting appropriate kernels and parameters always result in wrong interpretation of knowledge. The case becomes much worse when the data contains multiple classes. The ability to select the optimal kernel has been a research since SVM-based classifiers developed. In data-centric applications like diagnosis of diseases, object recognition etc, the approach of kernel selection is important. The risk minimization and performance boosting are the basic objectives of multi-kernel learning. This paper focuses on multi-kernel implementation of SVM over multi-class data sets by adopting different algorithms for solving minimization issues and performance boosting. A series of data sets collected from different domains like protein structure study, molecular biology and disease diagnosis. Experimental study with implementation of new algorithms has generated and an efficient kernel recommendation strategy for each domain is chosen.

Key-Words / Index Term

How to cite this article
Kanchana .J , Gladston Raj .S , “Efficient Multi Kernel Implementation optimised with Weight Updation Algorithm”, Journal for Advance Research in Applied Sciences, 4, Issue-7, pp.201-211, Dec 2017. DPI:16.10089.JARAS.V4.I7.2239