Paper Abstract and Keywords |
Presentation |
2016-11-24 14:50
Hybrid PSO and GSA-based Band Prioritization for Band Selection of High Dimensional Data Sets Yang Lang Chang, Yung-Hao Lai, Ming-Xiu Xu, Jyh-Perng Fang (NTUT), Chihyuan Chu (G-AVE Technology Corporation) SANE2016-64 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
With the progress of remote sensing technology, the numbers of bands of hyperspectral images are increased rapidly. However, it leads to the increase of the computational complexity exponentially as problem size increases. The huge data quantity also causes the curse of dimensionality, which leads to the worse accuracy. Therefore, we must reduce the data quantity in order to prevent the curse of dimensionality. In this paper, we propose a novel method, which hybrids the particle swarm optimization (PSO) and gravitational search algorithms (GSA) with impurity function band prioritization (IFBP) method, to reduce and extract the number of bands of hyperspectral images. The experimental results show that the proposed approach can not only work efficiently for reducing the dimension of data sets but also having a better classification accuracy compared to other classifiers. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Particle Swarm Optimization / Gravitational Search Algorithm / Impurity Function Band Prioritization / Hyperspectral Images / Multi Criteria Decision Making / Analytic Hierarchy Process / / |
Reference Info. |
IEICE Tech. Rep., vol. 116, no. 319, SANE2016-64, pp. 67-70, Nov. 2016. |
Paper # |
SANE2016-64 |
Date of Issue |
2016-11-17 (SANE) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
SANE2016-64 |
|