30 October 2009 A novel multi-class imbalance method based on the SVDD
Author Affiliations +
Proceedings Volume 7498, MIPPR 2009: Remote Sensing and GIS Data Processing and Other Applications; 749856 (2009) https://doi.org/10.1117/12.832374
Event: Sixth International Symposium on Multispectral Image Processing and Pattern Recognition, 2009, Yichang, China
Abstract
In the application of machine learning, there is data imbalance problem generally. In view of multi-class imbalance particularities, in this paper a new method named Multi-SVDD is proposed. And it is presented aimed to solving the problem of multi-class imbalance based on the existing Support Vector Data Description (SVDD), which is used to solve multi-class learning questions and the solution method for two-class imbalance problem. It was verified on the UCI data set, and compared with multi-class classification method proposed. The experiment has indicated that the Multi-SVDD algorithm is efficient when it is used to solve the multi-class imbalanced classification problems.
© (2009) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Dong Cui, Dong Cui, Qiang Zhang, Qiang Zhang, Zhimin Miao, Zhimin Miao, } "A novel multi-class imbalance method based on the SVDD", Proc. SPIE 7498, MIPPR 2009: Remote Sensing and GIS Data Processing and Other Applications, 749856 (30 October 2009); doi: 10.1117/12.832374; https://doi.org/10.1117/12.832374
PROCEEDINGS
7 PAGES


SHARE
Back to Top