Structural regularization in quadratic logistic regression model

作者：

Highlights：

• C-GRESH model which achieve SH or WH principle is proposed.

• A fast algorithm is designed with theoretical guarantee its convergence.

• SCV-RIC is proposed and utilized to optimize the regularization parameters.

• Oracle property of adaptive C-GRESH is shown.

摘要

•C-GRESH model which achieve SH or WH principle is proposed.•A fast algorithm is designed with theoretical guarantee its convergence.•SCV-RIC is proposed and utilized to optimize the regularization parameters.•Oracle property of adaptive C-GRESH is shown.

论文关键词：Quadratic model,Heredity principle,Classification,C-GRESH,Oracle property

论文评审过程：Received 1 June 2018, Revised 6 September 2018, Accepted 7 October 2018, Available online 12 October 2018, Version of Record 21 November 2018.

论文官网地址：https://doi.org/10.1016/j.knosys.2018.10.012