Robust one-class SVM for fault detection
详细信息 查看全文
- 作者：Yingchao Xiao<sup>asup><sup> ; sup><sup>xiaoyc10@mails.tsinghua.edu.cn" class="auth_mail" title="E-mail the corresponding authorsup> ; Huangang Wang<sup>asup><sup> ; sup><sup>hgwang@tsinghua.edu.cn" class="auth_mail" title="E-mail the corresponding authorsup> ; Wenli Xu<sup>asup><sup> ; sup><sup>xuwl@tsinghua.edu.cn" class="auth_mail" title="E-mail the corresponding authorsup> ; Junwu Zhou<sup>bsup><sup> ; sup><sup>zhou_jw@bgrimm.com" class="auth_mail" title="E-mail the corresponding authorsup>
- 关键词：Robust to outliers ; One-class SVM ; One-class classification ; Fault detection ; Tennessee Eastman Process
- 刊名：Chemometrics and Intelligent Laboratory Systems
- 出版年：2016
- 出版时间：15 February 2016
- 年：2016
- 卷：151
- 期：Complete
- 页码：15-25
- 全文大小：1632 K
文摘
- ss="listitem" id="list_l0010">
- ss="label">•
A new method is proposed to reduce outliers' influence on OCSVM model.
- ss="label">•
Some suspected outliers are identified and removed first.
- ss="label">•
The decision boundary enclosing the cluster core of the target class is obtained.
- ss="label">•
The outliers are identified and removed based on this boundary.
- ss="label">•
The final OCSVM model is obtained on the remaining target samples.