On Consistency and Sparsity for Principal Components Analysis in High Dimensions.

On Consistency and Sparsity for Principal Components Analysis in High Dimensions.
复制标题

DOI:
10.1198/jasa.2009.0121
复制
发表时间:
2009-06-01
影响因子:
3.7
通讯作者:
Lu AY
Lu AY
中科院分区:
数学1区
文献类型:
--
作者:
Johnstone IM;Lu AY

文献摘要

参考文献

被引文献

相似文献

主成分分析(PCA)是一种经典的降维方法,用于对具有p个变量的向量的n个观测(或情况)的数据进行降维。当代数据集的p值通常与n相当,甚至比n大得多。在这种情况下,我们的主要主张是:(a)在对主模式应用任何PCA类型的搜索之前,需要进行一些初始降维,以及(B)通过在信号具有稀疏表示的基础上进行工作,可以最好地实现初始降维。我们描述了一个简单的渐近模型,其中通过标准PCA的领先主成分向量的估计是一致的,当且仅当p(n)/n→0。我们提供了一个简单的算法来选择一个子集的坐标与最大的样本方差,并表明,如果PCA是在选定的子集,然后恢复一致性,即使p(n)<$n。
Principal components analysis (PCA) is a classic method for the reduction of dimensionality of data in the form of n observations (or cases) of a vector with p variables. Contemporary datasets often have p comparable with or even much larger than n. Our main assertions, in such settings, are (a) that some initial reduction in dimensionality is desirable before applying any PCA-type search for principal modes, and (b) the initial reduction in dimensionality is best achieved by working in a basis in which the signals have a sparse representation. We describe a simple asymptotic model in which the estimate of the leading principal component vector via standard PCA is consistent if and only if p(n)/n→0. We provide a simple algorithm for selecting a subset of coordinates with largest sample variances, and show that if PCA is done on the selected subset, then consistency is recovered, even if p(n) ⪢ n.
DOI: 10.1137/050645506
发表时间: 2007-09-01
期刊: SIAM REVIEW
影响因子: 10.2
作者:
d'Aspremont, Alexandre;El Ghaoui, Laurent;Lanckriet, Gert R. G.
通讯作者: Lanckriet, Gert R. G.
DOI: 10.1088/0305-4470/27/6/015
发表时间: 1994-03-21
期刊: JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL
影响因子: --
作者:
BIEHL, M;MIETZNER, A
通讯作者: MIETZNER, A
DOI: 10.1214/08-aos618
发表时间: 2008-12-01
影响因子: 4.5
作者:
Nadler, Boaz
通讯作者: Nadler, Boaz
DOI: 10.1214/aoms/1177704248
发表时间: 1963-01-01
影响因子: --
作者:
ANDERSON, TW
通讯作者: ANDERSON, TW
DOI: 10.1088/0305-4470/29/13/021
发表时间: 1996-07-07
期刊: JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL
影响因子: --
作者:
Reimann, P;VandenBroeck, C;Bex, GJ
通讯作者: Bex, GJ