LP-SVR Model Selection Using an Inexact Globalized Quasi-Newton Strategy

• 作者: David Garcia Chaparro J. A. Perez Venzor Jose-Gerardo Rosiles Juan Cota-Ruiz Pablo Rivas-Perea
• 来源期刊: 智能学习系统与应用(英文)
• 学科: 医学
• 关键词: Hyper-Parameter Estimation Support VECTOR Regression Machine Learning Data Mining
• 年，卷（期）: 2013,（1）
• 页码范围: 19-28
• 页数: 10页
• 分类号: R73

摘要

In this paper we study the problem of model selection for a linear programming-based support vector machine for regression. We propose generalized method that is based on a quasi-Newton method that uses a globalization strategy and an inexact computation of first order information. We explore the case of two-class, multi-class, and regression problems. Simulation results among standard datasets suggest that the algorithm achieves insignificant variability when measuring residual statistical properties.