Safe feature screening rules for the regularized Huber regression
作者：
Highlights：
• Two safe feature screening rules are proposed for the regularized Huber regression.
• These rules can ensure that the dumped features have zero coefficients in solutions.
• The screening rules possess explicit expressions and low computational complexity.
• Experimental results show our rules significantly reduce the computational costs.
摘要
•Two safe feature screening rules are proposed for the regularized Huber regression.•These rules can ensure that the dumped features have zero coefficients in solutions.•The screening rules possess explicit expressions and low computational complexity.•Experimental results show our rules significantly reduce the computational costs.
论文关键词：Safe feature screening rules,High-dimensionality,Huber regression,Convex optimization,Duality theory
论文评审过程：Received 16 December 2019, Revised 24 June 2020, Accepted 28 June 2020, Available online 12 July 2020, Version of Record 12 July 2020.
论文官网地址：https://doi.org/10.1016/j.amc.2020.125500
