Fast Support Vector Data Description training using edge detection on large datasets

Chenlong Hu, Bo Zhou, Takayuki Furuzuki

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

Support Vector Data Description (SVDD) inherits properties of Support Vector Machines (SVM) and has become a prominent One Class Classifier (OCC). Same to standard SVM, its O (n3) time and O (n2) space complexities, where n is the number of training samples, have become major limitations in cases of large training datasets. As a simple and effective method, reducing the size of training dataset through reserving only samples mostly relevant to learned classifier, can be adopted to overcome the limitations. A trained SVDD enclosed decision boundary always locates on edge area of data distribution and is decided by a small subset of Support Vectors(SVs). Therefore, in this paper, we present a method based on edge detection such that edge samples mostly relevant to decision boundary can be preserved. And clustering techniques are also be applied to keep centroids representing the global distribution properties so as to avoid over-outside of decision boundary. To restrict the influences of noises, each training pattern is assigned with a weight. Experiments on real and artificial data sets prove that the classifier trained on reconstruction training set consisting of edge points and centroids can preserve performance with much faster training speed.

Original languageEnglish
Title of host publicationProceedings of the International Joint Conference on Neural Networks
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages2176-2182
Number of pages7
ISBN (Print)9781479914845
DOIs
Publication statusPublished - 2014 Sep 3
Event2014 International Joint Conference on Neural Networks, IJCNN 2014 - Beijing
Duration: 2014 Jul 62014 Jul 11

Other

Other2014 International Joint Conference on Neural Networks, IJCNN 2014
CityBeijing
Period14/7/614/7/11

Fingerprint

Data description
Edge detection
Classifiers
Support vector machines
Experiments

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence

Cite this

Hu, C., Zhou, B., & Furuzuki, T. (2014). Fast Support Vector Data Description training using edge detection on large datasets. In Proceedings of the International Joint Conference on Neural Networks (pp. 2176-2182). [6889718] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IJCNN.2014.6889718

Fast Support Vector Data Description training using edge detection on large datasets. / Hu, Chenlong; Zhou, Bo; Furuzuki, Takayuki.

Proceedings of the International Joint Conference on Neural Networks. Institute of Electrical and Electronics Engineers Inc., 2014. p. 2176-2182 6889718.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Hu, C, Zhou, B & Furuzuki, T 2014, Fast Support Vector Data Description training using edge detection on large datasets. in Proceedings of the International Joint Conference on Neural Networks., 6889718, Institute of Electrical and Electronics Engineers Inc., pp. 2176-2182, 2014 International Joint Conference on Neural Networks, IJCNN 2014, Beijing, 14/7/6. https://doi.org/10.1109/IJCNN.2014.6889718
Hu C, Zhou B, Furuzuki T. Fast Support Vector Data Description training using edge detection on large datasets. In Proceedings of the International Joint Conference on Neural Networks. Institute of Electrical and Electronics Engineers Inc. 2014. p. 2176-2182. 6889718 https://doi.org/10.1109/IJCNN.2014.6889718
Hu, Chenlong ; Zhou, Bo ; Furuzuki, Takayuki. / Fast Support Vector Data Description training using edge detection on large datasets. Proceedings of the International Joint Conference on Neural Networks. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 2176-2182
@inproceedings{823ab462a2424d829a4e76fd0031b08d,
title = "Fast Support Vector Data Description training using edge detection on large datasets",
abstract = "Support Vector Data Description (SVDD) inherits properties of Support Vector Machines (SVM) and has become a prominent One Class Classifier (OCC). Same to standard SVM, its O (n3) time and O (n2) space complexities, where n is the number of training samples, have become major limitations in cases of large training datasets. As a simple and effective method, reducing the size of training dataset through reserving only samples mostly relevant to learned classifier, can be adopted to overcome the limitations. A trained SVDD enclosed decision boundary always locates on edge area of data distribution and is decided by a small subset of Support Vectors(SVs). Therefore, in this paper, we present a method based on edge detection such that edge samples mostly relevant to decision boundary can be preserved. And clustering techniques are also be applied to keep centroids representing the global distribution properties so as to avoid over-outside of decision boundary. To restrict the influences of noises, each training pattern is assigned with a weight. Experiments on real and artificial data sets prove that the classifier trained on reconstruction training set consisting of edge points and centroids can preserve performance with much faster training speed.",
author = "Chenlong Hu and Bo Zhou and Takayuki Furuzuki",
year = "2014",
month = "9",
day = "3",
doi = "10.1109/IJCNN.2014.6889718",
language = "English",
isbn = "9781479914845",
pages = "2176--2182",
booktitle = "Proceedings of the International Joint Conference on Neural Networks",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Fast Support Vector Data Description training using edge detection on large datasets

AU - Hu, Chenlong

AU - Zhou, Bo

AU - Furuzuki, Takayuki

PY - 2014/9/3

Y1 - 2014/9/3

N2 - Support Vector Data Description (SVDD) inherits properties of Support Vector Machines (SVM) and has become a prominent One Class Classifier (OCC). Same to standard SVM, its O (n3) time and O (n2) space complexities, where n is the number of training samples, have become major limitations in cases of large training datasets. As a simple and effective method, reducing the size of training dataset through reserving only samples mostly relevant to learned classifier, can be adopted to overcome the limitations. A trained SVDD enclosed decision boundary always locates on edge area of data distribution and is decided by a small subset of Support Vectors(SVs). Therefore, in this paper, we present a method based on edge detection such that edge samples mostly relevant to decision boundary can be preserved. And clustering techniques are also be applied to keep centroids representing the global distribution properties so as to avoid over-outside of decision boundary. To restrict the influences of noises, each training pattern is assigned with a weight. Experiments on real and artificial data sets prove that the classifier trained on reconstruction training set consisting of edge points and centroids can preserve performance with much faster training speed.

AB - Support Vector Data Description (SVDD) inherits properties of Support Vector Machines (SVM) and has become a prominent One Class Classifier (OCC). Same to standard SVM, its O (n3) time and O (n2) space complexities, where n is the number of training samples, have become major limitations in cases of large training datasets. As a simple and effective method, reducing the size of training dataset through reserving only samples mostly relevant to learned classifier, can be adopted to overcome the limitations. A trained SVDD enclosed decision boundary always locates on edge area of data distribution and is decided by a small subset of Support Vectors(SVs). Therefore, in this paper, we present a method based on edge detection such that edge samples mostly relevant to decision boundary can be preserved. And clustering techniques are also be applied to keep centroids representing the global distribution properties so as to avoid over-outside of decision boundary. To restrict the influences of noises, each training pattern is assigned with a weight. Experiments on real and artificial data sets prove that the classifier trained on reconstruction training set consisting of edge points and centroids can preserve performance with much faster training speed.

UR - http://www.scopus.com/inward/record.url?scp=84908471854&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84908471854&partnerID=8YFLogxK

U2 - 10.1109/IJCNN.2014.6889718

DO - 10.1109/IJCNN.2014.6889718

M3 - Conference contribution

AN - SCOPUS:84908471854

SN - 9781479914845

SP - 2176

EP - 2182

BT - Proceedings of the International Joint Conference on Neural Networks

PB - Institute of Electrical and Electronics Engineers Inc.

ER -