摘要
针对Lasso问题,基于邻近梯度法,考虑在光滑项梯度及邻近算子的计算中存在误差,提出了一种带有误差项的不精确邻近梯度算法,并在适当条件下给出该算法的收敛速度分析。数值实验结果表明,不精确邻近梯度算法求解Lasso问题是有效的。
In order to solve the Lasso problem,based on the proximal-gradient algorithm,by considering the errors in the calculation of the gradient of the smooth term and in the proximity operator,an inexact proximal-gradient algorithm with errors is proposed,and analyse the convergence rate analysis of the algorithm under mild conditions.Numerical experiments show that the inexact proximal-algorithm is effective for solving Lasso problem.
作者
谢秋玲
徐宇淼
胡清洁
XIE Qiuling;XU Yumiao;HU Qingjie(School of Mathematics and Computing Science,Guilin University of Electronic Technology,Guilin 541004,China)
出处
《桂林电子科技大学学报》
2021年第3期236-240,共5页
Journal of Guilin University of Electronic Technology
基金
国家自然科学基金(11961011)。
关键词
Lasso问题
不精确邻近算子
邻近梯度法
Lasso problem
inexact proximity operator
inexact proximal-gradient algorithm