Improving support vector classification efficiency with principal component analysis

Conference proceedings article

Authors/Editors

Strategic Research Themes

Publication Details

Author list: Yaicharoen, Auapong; Yamada, Kuo;

Publisher: Elsevier

Publication year: 2021

Start page: 862

End page: 865

Number of pages: 4

ISBN: 9780738111278

ISSN: 0928-4931

eISSN: 1873-0191

URL: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85112802331&doi=10.1109%2fECTI-CON51831.2021.9454883&partnerID=40&md5=a09757df03329cf55c32e559c116fdc3

Languages: English-Great Britain (EN-GB)

View on publisher site

Abstract

In this paper, a combination of principal component analysis (PCA) technique and a support vector classification (SVC) technique to improve the speed of creating a data classifier is proposed. The dimensionality reduction using PCA technique is applied on a set of data before the process of finding support vectors to see if they can be used together to reduce calculation time when performing the binary classification on two linear separable data clusters. The proposed method is tested on several randomly generated data sets and compared two important metrics, time and accuracy, with traditional support vector classification technique. The results of the proposed method show speed improvement in some cases, comparing to using only the original SVC technique. The values obtained from the accuracy metric show that results from the proposed method does not suffer from the reduction of features in the test data sets. © 2021 IEEE.

Keywords

Support vector classification