Robust twin extreme learning machines with correntropy-based metric
作者：
Highlights：
• Based on correntropy and Laplacian kernel, a robust distance metric is proposed. A new non-convex fraction loss function is developed. Applying to TELM, a robust classification framework is proposed.
• The proposed metric includes and extends the traditional metrics and fractional loss function is a powerful adaptive cost in the presence of noise.
• An efficient optimization method is proposed to solve the model.
• Numerical experiments show that the proposed LCFTELM is effective and more robust to outliers.
摘要
•Based on correntropy and Laplacian kernel, a robust distance metric is proposed. A new non-convex fraction loss function is developed. Applying to TELM, a robust classification framework is proposed.•The proposed metric includes and extends the traditional metrics and fractional loss function is a powerful adaptive cost in the presence of noise.•An efficient optimization method is proposed to solve the model.•Numerical experiments show that the proposed LCFTELM is effective and more robust to outliers.
论文关键词：Robustness,Metric learning,Non-convex loss,Classification,Iterative algorithm
论文评审过程：Received 5 June 2020, Revised 20 November 2020, Accepted 17 December 2020, Available online 28 December 2020, Version of Record 2 January 2021.
论文官网地址：https://doi.org/10.1016/j.knosys.2020.106707
