| 在线分类算法收敛性分析 |
| |
| 引用本文： | 田明党,盛宝怀.在线分类算法收敛性分析[J].绍兴文理学院学报,2010,30(10). |
| |
| 作者姓名： | 田明党 盛宝怀 |
| |
| 作者单位： | 1. 宁波大学理学院,浙江宁波,315211
2. 绍兴文理学院数学系,浙江绍兴,312000 |
| |
| 摘 要： | 文章研究基于凸损失函数的系数正则化在线算法误差分析，给出了欧式空间上的一种不依赖于样本空间容量的算法，并且根据步长的选择给出了相应的学习速率和错分类误差．
|
| 关 键 词： | 分类算法 正则样本误差 正则系数 误差分析 |
Convergence Analysis of Online Classification Algorithms |
| |
| Tian Mingdang,Sheng Baohuai.Convergence Analysis of Online Classification Algorithms[J].Journal of Shaoxing College of Arts and Sciences,2010,30(10). |
| |
| Authors: | Tian Mingdang Sheng Baohuai |
| |
| Abstract: | The present paper deals with the error analysis of coefficient regularized online classification algorithms associated with convex loss functions. A novel capacity - independent approach based on the Eueliean Space is presented. Explicit learning rates and misclassification errors are also given for particular step sizes. |
| |
| Keywords: | classification algorithm regularized sample error coefficient regularization error analysis |
| 本文献已被 万方数据 等数据库收录！ |
