SVMTorch: Support vector machines for large-scale regression problems
SVMTorch: Support vector machines for large-scale regression problems
复制标题
DOI：
10.1162/15324430152733142
复制
发表时间：
2001-03-01
影响因子：
6
通讯作者：
Bengio, S
中科院分区：
文献类型：
--
作者：
Collobert, R;Bengio, S
用于回归问题的支持向量机（svm）是通过求解一个二次优化问题来训练的，该问题需要l(2)阶的内存和时间资源来求解，其中l为训练样例的数量。在本文中，我们提出了一种分解算法SVMTorch(1)，该算法类似于Joachims（1999）针对分类问题提出的SVM-Light，但适用于回归问题。使用该算法，可以有效地解决大规模回归问题（超过20000个示例）。Nodelib是Flake和Lawrence（2000）针对大规模回归问题提出的另一种公开可用的支持向量机算法，与Nodelib相比，时间得到了显著改善。最后，基于Lin（2000）最近的一篇论文，我们证明了我们的算法存在收敛性证明。
Support Vector Machines (SVMs) for regression problems are trained by solving a quadratic optimization problem which needs on the-order of l(2) memory and time resources to solve, where l is the number of training examples. In this paper, we propose a decomposition algorithm, SVMTorch(1), which is similar to SVM-Light proposed by Joachims (1999) for classification problems, but adapted to regression problems. With this algorithm, one can now efficiently solve large-scale regression problems (more than 20000 examples). Comparisons with Nodelib, another publicly available SVM algorithm for large-scale regression problems from Flake and Lawrence (2000) yielded significant time improvements. Finally, based on a recent paper from Lin (2000), we show that a convergence proof exists for our algorithm.