Learning Rate of Least Square Regressions with Some Kind of Mercer Kernel

摘要：

We consider the error estimate of least square regression with data dependent hypothesis and coecient regularization algorithms based on general kernel. When the kernel belongs to some kind of Mercer kernel, under a mild regularity condition on the regression function, we derive a dimensional-free learning rate 1/6 m. .

关键词： Square Regressions Data Dependent Hypothesis Coecient Regularization Mercer Kernel Learning Rate Introduction

作者: Sheng Baohui Duan Liqin Ye Peixin

作者单位: Department of Mathematics, Shaoxing College of Arts and Sciences, Shaoxing, Zhejiang 312000, China Institute of Mathematics, Hangzhou Dianzi University, Hangzhou 310018, China School of Mathematics and LPMC, Nankai University, Tianjin 300071, China

会议类型: 国际会议

会议名称: 2012 International Conference on Intelligent System Design and Engineering Applications(2012年智能系统设计与工程应用国际会议 ISDEA 2012)

会议地点: 三亚

会议语种:英文

页码: 329-332

在线出版日期: 2012-01-06（万方平台首次上网日期，不代表论文的发表时间）