Maximal Margin Support Vector Machine with Magnetic Field Effect
-
摘要: 为了提升模式分类泛化性能，该文提出一种新颖的具有磁场效应的大间隔支持向量机(MFSVM)。为了学习最优分割超平面，MFSVM通过引入最小化的q-磁场带，使得一类(或正常类)被包含其中，而另一类(或异常类)与该q-磁场带的间隔尽可能地大，从而实现类内内聚性的提高和类间间隔的增大，增强SVM学习泛化能力。在人造和实际数据集上实验结果显示，MFSVM分别在二类和一类模式分类上的性能均优于或等同于相关方法。Abstract: In this paper, a novel maximal margin Support Vector Machine with Magnetic Field effect (MFSVM) is proposed in allusion to the improvement of the generalization performance of pattern classification issue. By introducing a minimum q-magnetic field tape, the basic idea of MFSVM is to find an optimal hyper-plane with magnetic field effect such that one class (or normal patterns) can be enclosed in the q-magnetic field tape due to the magnetic attractive effect, while at the same time the margin between the q-magnetic field tape and the other class (or abnormal patterns) is as large as possible due to magnetic repulsion, thus implementing both maximum between-class margin and minimum within-class volume so as to improve the generalization capability of the proposed method. Experimental results obtained with benchmarking and synthetic datasets show that the proposed algorithm is effective and competitive to other related methods in such cases as two-class and one-class pattern classification respectively.
-
Chung Fu-Lai, Deng Zhao-hong, and Wang Shi-tong. From minimum enclosing ball to fast fuzzy inference system training on large datasets [J].IEEE Transactions on Fuzzy System.2009, 17(1):173-184[7]Wang Di, Zhang Bo, and Zhang Peng, et al.An online core vector machine with adaptive MEB adjustment [J].Pattern Recognition.2010, 43(1):3468-3482[8]Mller, K R, Mika S, Ratsch G, Tsuda K, and Schlkopf B. An introduction to kernel-based learning algorithms [J].IEEE Transactions on Neural Networks.2001, 12(2):181-201[9]Orabona F, Castellini C, and Caputo B, et al.On-line independent support vector machines [J].Pattern Recognition.2010, 43(1):1402-1412[10]Schlkopf B, Smola A J, Williamson R, and Bartlett P L. New support vector algorithms [J].Neural Computation.2000, 12(5):1207-1245[11]Tax D M J and Duin R P W. Support vector data description [J].Machine Learning.2004, 54(1):45-66[12]Wu Mingrui and Ye Jieping. A small sphere and large margin approach for novelty detection using training data with outliers [J].IEEE Transactions on Pattern Analysis and Machine Intelligence.2009, 31(11):2088-2092[13]邓乃扬, 田英杰. 支持向量机理论,算法与拓展[M]. 北京: 科学出版社, 2009: 5-180.[1]邓乃扬,田英杰.支持向量机理论,算法与拓展.北京:科学出版社,2009:5-180. -
计量
- 文章访问数: 3034
- HTML全文浏览量: 80
- PDF下载量: 665
- 被引次数: 0
下载:
