Multicategory ψ-learning and support vector machine:: Computational tools
Multicategory ψ-learning and support vector machine:: Computational tools
复制标题
DOI：
10.1198/106186005x37238
复制
发表时间：
2005-03-01
影响因子：
2.4
通讯作者：
Doss, H
中科院分区：
文献类型：
--
作者：
Liu, YF;Shen, XT;Doss, H
许多基于边缘的二进制分类技术，如支持向量机和PSI-学习，都提供了高性能。早些时候的一篇文章提出了一种新的多类别psi学习方法，该方法在泛化能力方面表现出很大的前景。然而，PSI学习在计算上是困难的，因为它需要处理一个非凸极小化问题。在本文中，我们提出了两种用于多类别PSI学习的计算工具。第一个是基于华盛顿特区的。第二种算法采用外逼近法，通过序列凹极小化得到全局极小值。数值算例表明，所提出的算法具有较好的性能。
Many margin-based binary classification techniques such as support vector machine (SVM) and psi-learning deliver high performance. An earlier article proposed a new multicategory psi-learning methodology that shows great promise in generalization ability. However, psi-learning is computationally difficult because it requires handling a nonconvex minimization problem. In this article, we propose two computational tools for multicategory psi-learning. The first one is based on d.c. algorithms and solved by sequential quadratic programming, while the second one uses the outer approximation method, which yields the global minimizer via sequential concave minimization. Numerical examples show the proposed algorithms perform well.