Robust twin support vector regression based on rescaled Hinge loss

作者：

Highlights：

• We propose to use rescaled Hinge loss function for Twin Support Vector Regression with regularizer and name it as Res-TSVR.

• We provide the dual formulations of the corresponding optimization problem in Res-TSVR.

• We give the analytic convergence proof of Res-TSVR.

• Res-TSVR is shown to be robust towards Gaussian and non-Gaussian noise.

摘要

•We propose to use rescaled Hinge loss function for Twin Support Vector Regression with regularizer and name it as Res-TSVR.•We provide the dual formulations of the corresponding optimization problem in Res-TSVR.•We give the analytic convergence proof of Res-TSVR.•Res-TSVR is shown to be robust towards Gaussian and non-Gaussian noise.

论文关键词：Twin support vector regression,Correntropy,Gaussian noise,Outliers,Linear kernel,Non-linear kernels,Res-TSVR

论文评审过程：Received 8 August 2019, Revised 5 March 2020, Accepted 22 April 2020, Available online 28 April 2020, Version of Record 3 May 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107395