An Unbalanced Penalty SVM for Fault Identification of BOSS

DOI

10.2991/msie-13.2013.95How to use a DOI?

Keywords

Fault Identification, Support Vector Machine, BOSS, Penalty Coefficient, Unbalanced Samples

Abstract

In order to solve classification error problems of support vector machine, which was used in the telecommunication business supporting system (BOSS), caused by the unbalanced ratio of positive samples, which stand for proper states of BOSS, and negative samples, which stand for the improper states, an unbalanced penalty SVM algorithm was proposed. In the proposed algorithm, values of the penalties were inverse to the ratio of the numbers of positive and negative samples, which means that the number of samples is higher, the lower the penalty coefficient. At last, in order to prove the effectiveness of the proposed algorithm, an experiment was conducted on the classification of running data of BOSS. The result of the experiment proved that the proposed SVM algorithm greatly reduces the negative sample misclassification when the ratio of positive and negative samples was not balanced, which proved the validity of the proposed algorithm.

Copyright

© 2013, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Zhifeng Chen
AU  - Minjing Peng
AU  - Bo Li
PY  - 2013/11
DA  - 2013/11
TI  - An Unbalanced Penalty SVM for Fault Identification of BOSS
BT  - Proceedings of the 2nd International Conference on Management Science and Industrial Engineering
PB  - Atlantis Press
SP  - 447
EP  - 450
SN  - 1951-6851
UR  - https://doi.org/10.2991/msie-13.2013.95
DO  - 10.2991/msie-13.2013.95
ID  - Chen2013/11
ER  -