Classification of sparse high-dimensional vectors

Classification of sparse high-dimensional vectors
复制标题

稀疏高维向量的分类

DOI:
10.1098/rsta.2009.0156
复制
发表时间:
2009
期刊:
Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences
影响因子:
--
通讯作者:
A. Tsybakov
A. Tsybakov
中科院分区:
--
文献类型:
--
作者:
Yu. I. Ingster;C. Pouet;A. Tsybakov

文献摘要

参考文献

被引文献

相似文献

我们研究基于大小为 m 的训练样本将 d 维向量分类为两类(其中一类是“纯噪声”)的问题。主要特点是尺寸 d 可以非常大。我们假设总体分布与噪声分布之间的差异仅在于平移,这是一个稀疏向量。对于高斯噪声,固定样本量m,维度d趋于无穷大,我们得到清晰的分类边界,即分类成功的可能性的充要条件。我们提出达到这个边界的分类器。我们还将结果扩展到样本大小 m 取决于 d 并满足条件 0 ≤ γ < 1 的情况,以及满足 Cramér 条件的非高斯噪声的情况。
We study the problem of classification of d-dimensional vectors into two classes (one of which is ‘pure noise’) based on a training sample of size m. The main specific feature is that the dimension d can be very large. We suppose that the difference between the distribution of the population and that of the noise is only in a shift, which is a sparse vector. For Gaussian noise, fixed sample size m, and dimension d that tends to infinity, we obtain the sharp classification boundary, i.e. the necessary and sufficient conditions for the possibility of successful classification. We propose classifiers attaining this boundary. We also give extensions of the result to the case where the sample size m depends on d and satisfies the condition , 0 ≤ γ < 1, and to the case of non-Gaussian noise satisfying the Cramér condition.
DOI: 10.1016/s1535-6108(02)00032-6
发表时间: 2002-03-01
期刊: CANCER CELL
影响因子: 50.3
作者:
Yeoh, EJ;Ross, ME;Downing, JR
通讯作者: Downing, JR
DOI: 10.1182/blood-2004-03-1154
发表时间: 2004-12-01
期刊: BLOOD
影响因子: 20.3
作者:
Ross, ME;Mahfouz, R;Downing, JR
通讯作者: Downing, JR
DOI: 10.1182/blood-2003-01-0338
发表时间: 2003-10-15
期刊: BLOOD
影响因子: 20.3
作者:
Ross, ME;Zhou, XD;Downing, JR
通讯作者: Downing, JR