Multi-Class Support Vector Machine via Maximizing Multi-Class Margins
Multi-Class Support Vector Machine via Maximizing Multi-Class Margins
复制标题
DOI：
10.24963/ijcai.2017/440
复制
发表时间：
2017-08
期刊：
影响因子：
--
通讯作者：
Jie Xu;Xianglong Liu;Zhouyuan Huo;Cheng Deng;F. Nie;Heng Huang
中科院分区：
文献类型：
--
作者：
Jie Xu;Xianglong Liu;Zhouyuan Huo;Cheng Deng;F. Nie;Heng Huang
支持向量机（SVM）最初是作为一种二值分类模型提出的，并在许多应用中取得了巨大的成功。在现实中，更常见的是解决一个有两个以上类的问题。因此，将支持向量机扩展到多类分类器是很自然的。基于二值支持向量机构建多类分类器的研究有很多，如一对一策略（OvsR）、一对一策略（OvsO）和Weston的多类支持向量机。前两种方法将多类问题拆分为多个二分类子问题，我们需要训练多个二分类器。Weston的多类SVM是通过保证风险约束并施加特定的正则化（如Frobenius范数）而形成的。它不是通过最大化超平面与训练数据之间的余量来推导的，而这是支持向量机的动机。本文从训练点与超平面的边界最大化的角度出发，提出了一种多类SVM模型，并分析了该模型与其他相关方法的关系。实验表明，与其他相关方法相比，我们的模型可以得到更好或比较好的结果。
Support Vector Machine (SVM) is originally proposed as a binary classification model with achieving great success in many applications. In reality, it is more often to solve a problem which has more than two classes. So, it is natural to extend SVM to a multi-class classifier. There have been many works proposed to construct a multi-class classifier based on binary SVM, such as one versus rest strategy (OvsR), one versus one strategy (OvsO) and Weston’s multi-class SVM. The first two split the multi-class problem to multiple binary classification subproblems, and we need to train multiple binary classifiers. Weston’s multi-class SVM is formed by ensuring risk constraints and imposing a specific regularization, like Frobenius norm. It is not derived by maximizing the margin between hyperplane and training data which is the motivation in SVM. In this paper, we propose a multiclass SVM model from the perspective of maximizing margin between training points and hyperplane, and analyze the relation between our model and other related methods. In the experiment, it shows that our model can get better or compared results when comparing with other related methods.