A fast tracking algorithm for generalized LARS/LASSO

被引：20

作者：

Keerthi, S. Sathiya

Shevade, Shirish

机构：

[1] Yahoo Res, Burbank, CA 91504 USA

[2] Indian Inst Sci, Dept Comp Sci & Automat, Bangalore 560012, Karnataka, India

来源：

IEEE TRANSACTIONS ON NEURAL NETWORKS | 2007年 / 18卷 / 06期

关键词：

generalized Least Angle Regression (LARS); Least Absolute Shrinkage and Selection Operator (LASSO); sparse logistic regression;

D O I：

10.1109/TNN.2007.900229

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This letter gives an efficient algorithm for tracking the solution curve of sparse logistic regression With respect to the L I regularization parameter. The algorithm is based on approximating the logistic regression loss by a piecewise quadratic function, using Rosset and Zhu's path tracking algorithm on the approximate problem, and then applying a correction to get to the true path. Application of the algorithm to text classification and sparse kernel logistic regression shows that the algorithm is efficient.

引用

页码：1826 / 1830

页数：5

共 14 条

[1]

BACH FR, 2005, NEURAL INFORM PROCES, V17

[2] Least angle regression - Rejoinder [J].

Efron, B ;

Hastie, T ;

Johnstone, I ;

Tibshirani, R .

ANNALS OF STATISTICS, 2004, 32 (02) :494-499

[3]

Genkin Alexander, 2004, LARGE SCALE BAYESIAN

[4]

KEERTHI SS, 2005, P 22 INT C MACH LEAR, P417

[5]

Madigan D, 2004, ANN STAT, V32, P465

[6] On the LASSO and its dual [J].

Osborne, MR ;

Presnell, B ;

Turlach, BA .

JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS, 2000, 9 (02) :319-337

[7]

RATSCH TG, 1998, BENCHMARK REPOSITORY

[8]

ROSSET S, 2005, ADV NEURAL INFORM PR, V17

[9]

ROSSET S, 2004, PIECEWISE LINEAR REG

[10] The generalized LASSO [J].

Roth, V .

IEEE TRANSACTIONS ON NEURAL NETWORKS, 2004, 15 (01) :16-28

← 1 2 →