DMCA
Robust truncated-hinge-loss support vector machines (2006)
| Citations: | 11 - 5 self |
BibTeX
@MISC{Wu06robusttruncated-hinge-loss,
author = {Yichao Wu and Yufeng Liu},
title = {Robust truncated-hinge-loss support vector machines},
year = {2006}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Abstract. With its elegant margin theory and accurate classification perfor-mance, the Support Vector Machine (SVM) has been widely applied in both machine learning and statistics. Despite its success and popularity, it still has some drawbacks in certain situations. In particular, the SVM classifier can be very sensitive to outliers in the training sample. Moreover, the number of support vectors (SVs) can be very large in many applications. To solve these problems, [WL06] proposed a new SVM variant, the robust truncated-hinge-loss SVM (RSVM), which uses a truncated hinge loss. In this paper, we apply the operation of truncation on the multicategory hinge loss proposed by [LLW04]. We show that the proposed robust multicategory truncated-hinge-loss SVM (RMSVM) is more robust to outliers and deliver more accurate classifiers using a smaller set of SVs than the original multicategory SVM (MSVM) proposed by [LLW04]. 1.
Keyphrases
robust truncated-hinge-loss support vector machine accurate classifier support vector machine elegant margin theory many application truncated hinge loss robust multicategory truncated-hinge-loss svm accurate classification perfor-mance robust truncated-hinge-loss svm machine learning certain situation new svm variant multicategory hinge loss support vector svm classifier original multicategory svm training sample
