Normalized neighborhood component feature selection and feasible-improved weight allocation for input variable selection
作者：
Highlights：
• The nNCFS is proposed by normalizing scales between mean loss and regularization terms.
• Robustness and computational cost of the nNCFS are significantly improved.
• The FIWA method is proposed considering a design optimization problem.
• Feasible and improved designs are obtained using the FIWA method.
• The optimum designs are obtained using fewer significant input variables.
摘要
•The nNCFS is proposed by normalizing scales between mean loss and regularization terms.•Robustness and computational cost of the nNCFS are significantly improved.•The FIWA method is proposed considering a design optimization problem.•Feasible and improved designs are obtained using the FIWA method.•The optimum designs are obtained using fewer significant input variables.
论文关键词：Normalized neighborhood component feature selection,Feasible-improved weight allocation,Input variable selection,Multi-response system,Design optimization,Body-in-white
论文评审过程：Received 22 January 2020, Revised 11 January 2021, Accepted 9 February 2021, Available online 10 February 2021, Version of Record 2 March 2021.
论文官网地址：https://doi.org/10.1016/j.knosys.2021.106855
