An improved SVM-KM model for imbalanced datasets

Weiguo Deng; Li Wang; Yiyang Wang; Zhong Qian

doi:10.1109/ICICEE.2012.35

An improved SVM-KM model for imbalanced datasets

Weiguo Deng^*
, Li Wang
, Yiyang Wang
, Zhong Qian

^*Corresponding author for this work

School of Economics and Management

Beihang University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Scopus citations

Abstract

Support vector machine is a widely used machine learning technique. SVM-KM model can speed SVM training by eliminating non support vectors, but imbalanced datasets will affect the classification accuracy. In this paper, we proposed an improved SVM-KM model, which assign different error costs to different classes. Based on the simulation results, the improved SVM-KM model performed best for imbalanced datasets.

Original language	English
Title of host publication	Proceedings of the 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012
Pages	100-103
Number of pages	4
DOIs	https://doi.org/10.1109/ICICEE.2012.35
State	Published - 2012
Event	2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012 - Xi'an, China Duration: 23 Aug 2012 → 25 Aug 2012

Publication series

Name	Proceedings of the 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012

Conference

Conference	2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012
Country/Territory	China
City	Xi'an
Period	23/08/12 → 25/08/12

Keywords

different error costs
imbalanced datasets
k-means
support vector machine

Access to Document

10.1109/ICICEE.2012.35

Cite this

Deng, W., Wang, L., Wang, Y., & Qian, Z. (2012). An improved SVM-KM model for imbalanced datasets. In Proceedings of the 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012 (pp. 100-103). Article 6322324 (Proceedings of the 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012). https://doi.org/10.1109/ICICEE.2012.35

@inproceedings{08ca0831c18c4d92a50cc970476d5aac,

title = "An improved SVM-KM model for imbalanced datasets",

abstract = "Support vector machine is a widely used machine learning technique. SVM-KM model can speed SVM training by eliminating non support vectors, but imbalanced datasets will affect the classification accuracy. In this paper, we proposed an improved SVM-KM model, which assign different error costs to different classes. Based on the simulation results, the improved SVM-KM model performed best for imbalanced datasets.",

keywords = "different error costs, imbalanced datasets, k-means, support vector machine",

author = "Weiguo Deng and Li Wang and Yiyang Wang and Zhong Qian",

year = "2012",

doi = "10.1109/ICICEE.2012.35",

language = "英语",

isbn = "9780769547923",

series = "Proceedings of the 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012",

pages = "100--103",

booktitle = "Proceedings of the 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012",

note = "2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012 ; Conference date: 23-08-2012 Through 25-08-2012",

}

Deng, W, Wang, L, Wang, Y & Qian, Z 2012, An improved SVM-KM model for imbalanced datasets. in Proceedings of the 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012., 6322324, Proceedings of the 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012, pp. 100-103, 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012, Xi'an, China, 23/08/12. https://doi.org/10.1109/ICICEE.2012.35

An improved SVM-KM model for imbalanced datasets. / Deng, Weiguo; Wang, Li; Wang, Yiyang et al.
Proceedings of the 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012. 2012. p. 100-103 6322324 (Proceedings of the 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - An improved SVM-KM model for imbalanced datasets

AU - Deng, Weiguo

AU - Wang, Li

AU - Wang, Yiyang

AU - Qian, Zhong

PY - 2012

Y1 - 2012

N2 - Support vector machine is a widely used machine learning technique. SVM-KM model can speed SVM training by eliminating non support vectors, but imbalanced datasets will affect the classification accuracy. In this paper, we proposed an improved SVM-KM model, which assign different error costs to different classes. Based on the simulation results, the improved SVM-KM model performed best for imbalanced datasets.

AB - Support vector machine is a widely used machine learning technique. SVM-KM model can speed SVM training by eliminating non support vectors, but imbalanced datasets will affect the classification accuracy. In this paper, we proposed an improved SVM-KM model, which assign different error costs to different classes. Based on the simulation results, the improved SVM-KM model performed best for imbalanced datasets.

KW - different error costs

KW - imbalanced datasets

KW - k-means

KW - support vector machine

UR - https://www.scopus.com/pages/publications/84869992865

U2 - 10.1109/ICICEE.2012.35

DO - 10.1109/ICICEE.2012.35

M3 - 会议稿件

AN - SCOPUS:84869992865

SN - 9780769547923

T3 - Proceedings of the 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012

SP - 100

EP - 103

BT - Proceedings of the 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012

T2 - 2012 International Conference on Industrial Control and Electronics Engineering, ICICEE 2012

Y2 - 23 August 2012 through 25 August 2012

ER -

An improved SVM-KM model for imbalanced datasets

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this