By Topic

Knowledge Discovery Employing Grid Scheme Least Squares Support Vector Machines Based on Orthogonal Design Bee Colony Algorithm

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Tsung-Jung Hsieh ; Dept. of Ind. Eng. & Eng. Manage., Nat. Tsing Hua Univ., Hsinchu, Taiwan ; Wei-Chang Yeh

This paper proposes a concept for machine learning that integrates a grid scheme (GS) into a least squares support vector machine (LSSVM) (called GS-LSSVM) with a mixed kernel in order to solve data classification problems. The purpose of GS-LSSVM is to execute feature selections, mixed kernel applications, and parameter optimization in a learning paradigm. The proposed learning paradigm includes three steps. First, an orthogonal design is utilized to initialize the number of input features and candidate parameters stored in GS. Then, the features are randomly selected according to the first grid acquired from the first step. These features and the candidate parameters are then passed to LSSVM. Finally, an artificial bee colony algorithm, the recently popular heuristic algorithm, is used to optimize parameters for LSSVM learning. For illustration and evaluation purposes, ten remarkable data sets from the University of California Irvine database are used as testing targets. The experimental results reveal that the proposed GS-LSSVM can produce a classification model more easily interpreted using a small number of features. In terms of accuracy (hit ratio), the GS-LSSVM can significantly outperform other methods listed in this paper. These findings imply that the GS-LSSVM is a promising approach to classification exploration.

Published in:

Systems, Man, and Cybernetics, Part B: Cybernetics, IEEE Transactions on  (Volume:41 ,  Issue: 5 )