On generalization bounds, projection profile, and margin distribution

On generalization bounds, projection profile, and margin distribution
复制标题

DOI:
--
复制
发表时间:
2002-07
期刊:
--
影响因子:
--
通讯作者:
A. Garg;Sariel Har-Peled;D. Roth
A. Garg;Sariel Har-Peled;D. Roth
中科院分区:
其他
文献类型:
--
作者:
A. Garg;Sariel Har-Peled;D. Roth

文献摘要

被引文献

相似文献

我们研究了线性学习算法的泛化性质,并提出了一种依赖于数据的方法,用于推导依赖于边界分布的泛化界限。我们的方法使用随机投影技术,允许在有效的、较低的数据维度中使用现有的VC维度界限。我们的边界比现有的边界更紧,并且(有时)为现实世界的高维问题提供了信息性的泛化边界。
We study generalization properties of linear learning algorithms and develop a data dependent approach that is used to derive generalization bounds that depend on the margin distribution. Our method uses random projection techniques to allow the use of existing VC dimension bounds in the effective, lower, dimension of the data. Our bounds are tighter than existing bounds and (sometimes) give informative generalization bounds for real world, high dimensional problems.