| 摘要: |
| 由 Sch?lkopf 等人提出的ν支持向量回归机具有通过参数ν控制支持向量和错误向量个数的优点,然而与标准的支持向量机相比,其形式更为复杂,迄今为止仍没有有效的算法计算ν解路径.基于ν支持向量回归机的修改形式,提出了一种新的解路径算法,它能够追踪参数ν对应的所有解,并通过理论分析和实验,说明了该算法能够尽可能地避免不可行的更新路径,并在有限步内拟合出所有的ν解路径. |
| 关键词: 模型选择 解路径 ν支持向量回归机 支持向量机 机器学习 |
| DOI：10.3724/SP.J.1001.2012.04153 |
| 分类号: |
| 基金项目:国家自然科学基金重点项目(61139002); 国家自然科学基金青年科学基金(61202137); 江苏高校优势学科建设工程资助项目; 南京信息工程大学科研启动费(20110433) |
|
| Effective ν-Path Algorithm for ν-Support Vector Regression |
|
GU Bin1,2,3, WANG Jian-Dong3
|
|
1.Jiangsu Engineering Center of Network Monitoring (Nanjing University of Information Science and Technology), Nanjing 210044,China;2.School of Computer and Software, Nanjing University of Information Science and Technology, Nanjing 210044, China;3.Department of Computer Science and Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China
|
| Abstract: |
| The ν-support vector regression (ν-SVR) proposed by Sch lkopf, et al., has the advantage of using theparameter ν to control the number of support vectors and margin errors, however, compared to ε-SVR, itsformulation is more complicated. Until now, there have been no effective methods used to compute the ν-path for it.This paper proposes a new solution path algorithm, which is designed based on a modified formulation of ν-SVRand traces the solution path with respect to the parameter ν. Through theoretical analysis and experiments, resultscan show that the algorithm can avoid the infeasible updating path, and fit the entire ν-path in finite steps. |
| Key words: model selection solution path ν-support vector regression support vector machine machinelearning |
