arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2006.14062math.STcs.LGmath.PRstat.MLstat.TH
收藏

PCA的$\ell_p$理论与谱聚类

An $\ell_p$ theory of PCA and spectral clustering

EPFL(洛桑联邦理工学院);Princeton University(普林斯顿大学);Columbia University(哥伦比亚大学)

Emmanuel Abbe, Jianqing Fan, Kaizheng Wang

摘要:

主成分分析(Principal Component Analysis,PCA)是统计学和机器学习领域的一个重要工具。虽然现有的PCA研究主要集中在主成分及其相关的特征值的恢复,有几个精确的表征个别主成分得分,产生低维嵌入的样本。这阻碍了各种光谱方法的分析。在本文中,我们首先发展了一个$\ell_p$扰动理论的中空版本的主成分分析在Hilbert空间中,证明了改进的香草主成分分析在异方差噪声的存在。通过一种新的$\ell_p$分析的特征向量,我们调查的主成分得分向量的入口行为,并表明他们可以近似的Gram矩阵在$\ell_p$范数的线性泛函,其中包括$\ell_2$和$\ell_\infty$作为特殊的例子。对于亚高斯混合模型,给出最优边界的$p$的选择取决于信噪比,这进一步产生谱聚类的最优性保证。对于上下文社区检测,$\ell_p$理论导致一个简单的频谱算法,达到准确恢复的信息阈值。作为特殊情况,这些还为高斯混合模型和随机块模型提供了最佳恢复结果。

英文摘要:

Principal Component Analysis (PCA) is a powerful tool in statistics and machine learning. While existing study of PCA focuses on the recovery of principal components and their associated eigenvalues, there are few precise characterizations of individual principal component scores that yield low-dimensional embedding of samples. That hinders the analysis of various spectral methods. In this paper, we first develop an $\ell_p$ perturbation theory for a hollowed version of PCA in Hilbert spaces which provably improves upon the vanilla PCA in the presence of heteroscedastic noises. Through a novel $\ell_p$ analysis of eigenvectors, we investigate entrywise behaviors of principal component score vectors and show that they can be approximated by linear functionals of the Gram matrix in $\ell_p$ norm, which includes $\ell_2$ and $\ell_\infty$ as special examples. For sub-Gaussian mixture models, the choice of $p$ giving optimal bounds depends on the signal-to-noise ratio, which further yields optimality guarantees for spectral clustering. For contextual community detection, the $\ell_p$ theory leads to a simple spectral algorithm that achieves the information threshold for exact recovery. These also provide optimal recovery results for Gaussian mixture and stochastic block models as special cases.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑