Handling the impact of feature uncertainties on SVM: A robust approach based on Sobol sensitivity analysis

作者：

Highlights：

• A new approach is proposed to evaluate the impact of feature uncertainties on SVM.

• Sobol analysis is applied to quantify of the impact of each feature uncertainties on SVM.

• Feature weights based on Sobol indices are introduced to improve the SVM robustness.

摘要

•A new approach is proposed to evaluate the impact of feature uncertainties on SVM.•Sobol analysis is applied to quantify of the impact of each feature uncertainties on SVM.•Feature weights based on Sobol indices are introduced to improve the SVM robustness.

论文关键词：Support vector machines,Sobol sensitivity analysis,Uncertainty,Robust classification

论文评审过程：Received 25 November 2020, Revised 9 July 2021, Accepted 27 July 2021, Available online 19 October 2021, Version of Record 23 October 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115691