科学研究
报告题目:

Smoothing Newton Method for L0-L2 Regularized Least Square Problems

报告人:

吕锡亮 教授(武大2021欧洲杯买球平台官网)

报告时间:

报告地点:

2021欧洲杯买球平台官网东北楼四楼报告厅(404)

报告摘要:

In this talk, we consider a multi-parameter (L0-L2) regularization for the linear regression with sparse contraints. A necessary and sufficient condition for the coordinate-wise minimizer is a discontinuous inclusion. A smoothing Newton algorithm with linear search is proposed. The global convergence of this approach is proved, and several numerical examples are given to illustrate the effiency of the algorithm.