Speeding up L 2-loss support vector regression by random Fourier features

• Published in: Communications in statistics. Simulation and computation Vol. 53; no. 2; pp. 933 - 951
• Main Author: Zheng, Songfeng
• Format: Journal Article
• Language: English
• Published: Taylor & Francis 01.02.2024
• Subjects: Random Fourier features; Smooth approximation; Squared ϵ-insensitive loss function; Support vector regression; Weighted least square

To avoid the expensive quadratic programming in the L 2 -loss support vector regression (SVR) model, smooth approximation and iteratively reweighted least square (IRLS) techniques were introduced in literature, resulting in smoothed SVR (SSVR) and IRLS-SVR. However, for nonlinear models, SSVR and IR...