Abstract:
Support Vector Machine(SVM) is a new technique for data mining. It has wide applications in various fields and is a research hot pot of the machine learning field. Due to its differentiability, smooth SVM, a reformulation of standard SVM, has shown advantages in many aspects. This paper introduces the basic principle of smooth SVM, SSVM model, PSSVM model and a new class of smoothing functions, it focuses on the open and key problems in this field. It discusses some promising directions of smooth SVM research for future work.
Key words:
pattern recognition,
classification,
Support Vector Machine(SVM),
smooth
摘要: 支持向量机是数据挖掘的新方法,由于其显著的优点,因此得到了广泛的应用。光滑支持向量机是标准支持向量机的一种改进形式,其应用已显示出了优越性。该文论述光滑支持向量机(SSVM)的基本原理、SSVM 模型、多项式光滑的支持向量机模型以及一类新的光滑函数。讨论光滑支持向量机目前尚未解决的一些关键问题,并展望了今后的发展趋势,为进一步研究光滑支持向量机理论提供了基本思路。
关键词:
模式识别,
分类,
支持向量机,
光滑
CLC Number:
XIONG Jin-zhi; HU Jin-lian; YUAN Hua-qiang. Principles and Advances of Smooth Support Vector Machine[J]. Computer Engineering, 2008, 34(13): 172-173,.
熊金志;胡金莲;袁华强. 光滑支持向量机的原理和进展[J]. 计算机工程, 2008, 34(13): 172-173,.