Analysis of regularized least squares ranking with centered reproducing kernel

被引:1
作者
He, Fangchao [1 ]
Zheng, Lie [1 ]
机构
[1] Hubei Univ Technol, Sch Sci, Wuhan 430068, Peoples R China
关键词
Ranking; regularization; integral operator; centered reproducing kernel; GENERALIZATION BOUNDS; LEARNING-THEORY; ALGORITHMS; BIAS;
D O I
10.1142/S0219691323500613
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
The existing results of the regularized least squares ranking (RLSR) algorithm are mainly related to the reproducing kernels. In this paper, we go beyond this limitation by investigating the convergence rates of ranking with centered reproducing kernel (CRK). We develop a new error analysis framework by means of CRK and integral operator. The result shows that the convergence rate of RLSR is still optimal as those in the traditional analysis framework.
引用
收藏
页数:26
相关论文
共 35 条
[1]  
Agarwal S, 2005, J MACH LEARN RES, V6, P393
[2]  
Agarwal S, 2009, J MACH LEARN RES, V10, P441
[3]   THEORY OF REPRODUCING KERNELS [J].
ARONSZAJN, N .
TRANSACTIONS OF THE AMERICAN MATHEMATICAL SOCIETY, 1950, 68 (MAY) :337-404
[4]  
Blanchard G, 2010, Arxiv, DOI arXiv:1009.5839
[5]   Debiased magnitude-preserving ranking: Learning rate and bias characterization [J].
Chen, Hong ;
Wang, Yingjie ;
Song, Biqin ;
Li, Han .
JOURNAL OF MATHEMATICAL ANALYSIS AND APPLICATIONS, 2020, 486 (02)
[6]   The convergence rate of a regularized ranking algorithm [J].
Chen, Hong .
JOURNAL OF APPROXIMATION THEORY, 2012, 164 (12) :1513-1519
[7]  
Clemencon S, 2007, J MACH LEARN RES, V8, P2671
[8]   Learning to order things [J].
Cohen, WW ;
Schapire, RE ;
Singer, Y .
JOURNAL OF ARTIFICIAL INTELLIGENCE RESEARCH, 1999, 10 :243-270
[9]  
Cortes C, 2012, J MACH LEARN RES, V13, P795
[10]   Best choices for regularization parameters in learning theory: On the bias-variance problem [J].
Cucker, F ;
Smale, S .
FOUNDATIONS OF COMPUTATIONAL MATHEMATICS, 2002, 2 (04) :413-428