Kernel spectral clustering of large dimensional data

This article proposes a first analysis of kernel spectral clustering methods in the regime where the dimension $p$ of the data vectors to be clustered and their number $n$ grow large at the same rate. We demonstrate, under a $k$-class Gaussian mixture model, that the normalized Laplacian matrix associated with the kernel matrix asymptotically behaves similar to a so-called spiked random matrix. Some of the isolated eigenvalue-eigenvector pairs in this model are shown to carry the clustering information upon a separability condition classical in spiked matrix models. We evaluate precisely the position of these eigenvalues and the content of the eigenvectors, which unveil important (sometimes quite disruptive) aspects of kernel spectral clustering both from a theoretical and practical standpoints. Our results are then compared to the actual clustering performance of images from the MNIST database, thereby revealing an important match between theory and practice.

[1]  V. Marčenko,et al.  DISTRIBUTION OF EIGENVALUES FOR SOME SETS OF RANDOM MATRICES , 1967 .

[2]  F. T. Wright,et al.  A Bound on Tail Probabilities for Quadratic Forms in Independent Random Variables , 1971 .

[3]  Charles R. Johnson,et al.  Matrix analysis , 1985, Statistical Inference for Engineers and Data Scientists.

[4]  J. W. Silverstein,et al.  Analysis of the limiting spectral distribution of large dimensional random matrices , 1995 .

[5]  J. W. Silverstein,et al.  On the empirical distribution of eigenvalues of a class of large dimensional random matrices , 1995 .

[6]  J. W. Silverstein,et al.  No eigenvalues outside the support of the limiting spectral distribution of large-dimensional sample covariance matrices , 1998 .

[7]  I. Johnstone On the distribution of the largest principal component , 2000 .

[8]  Michael I. Jordan,et al.  On Spectral Clustering: Analysis and an algorithm , 2001, NIPS.

[9]  S. Péché,et al.  Phase transition of the largest eigenvalue for nonnull complex sample covariance matrices , 2004, math/0403022.

[10]  Yann LeCun,et al.  The mnist database of handwritten digits , 2005 .

[11]  Christopher M. Bishop,et al.  Pattern Recognition and Machine Learning (Information Science and Statistics) , 2006 .

[12]  Nasser M. Nasrabadi,et al.  Pattern Recognition and Machine Learning , 2006, Technometrics.

[13]  D. Paul ASYMPTOTICS OF SAMPLE EIGENSTRUCTURE FOR A LARGE DIMENSIONAL SPIKED COVARIANCE MODEL , 2007 .

[14]  Ulrike von Luxburg,et al.  A tutorial on spectral clustering , 2007, Stat. Comput..

[15]  Mikhail Belkin,et al.  Consistency of spectral clustering , 2008, 0804.0678.

[16]  R. Couillet,et al.  Large System Analysis of Linear Precoding in MISO Broadcast Channels with Limited Feedback , 2009 .

[17]  Noureddine El Karoui,et al.  Concentration of measure and spectra of random matrices: Applications to correlation matrices, elliptical distributions and beyond , 2009, 0912.1950.

[18]  Raj Rao Nadakuditi,et al.  The eigenvalues and eigenvectors of finite, low rank perturbations of large random matrices , 2009, 0910.2120.

[19]  Noureddine El Karoui,et al.  The spectrum of kernel random matrices , 2010, 1001.0492.

[20]  A. Guionnet,et al.  Large deviations of the extreme eigenvalues of random deformations of matrices , 2010, Probability Theory and Related Fields.

[21]  A. Guionnet,et al.  Fluctuations of the Extreme Eigenvalues of Finite Rank Deformations of Random Matrices , 2010, 1009.0145.

[22]  Mérouane Debbah,et al.  A Deterministic Equivalent for the Analysis of Correlated MIMO Multiple Access Channels , 2009, IEEE Transactions on Information Theory.

[23]  L. Pastur,et al.  Eigenvalue Distribution of Large Random Matrices , 2011 .

[24]  L. Erdős Universality of Wigner random matrices: a survey of recent results , 2010, 1004.0861.

[25]  Walid Hachem,et al.  The outliers among the singular values of large rectangular random matrices with additive fixed rank deformation , 2012, 1207.0471.

[26]  Raj Rao Nadakuditi,et al.  The singular values and vectors of low rank perturbations of large rectangular random matrices , 2011, J. Multivar. Anal..

[27]  R. Couillet,et al.  Analysis of the limiting spectral measure of large random matrices of the separable covariance type , 2013, 1310.8094.

[28]  Walid Hachem,et al.  Fluctuations of Spiked Random Matrix Models and Failure Diagnosis in Sensor Networks , 2011, IEEE Transactions on Information Theory.

[29]  Philippe Loubaton,et al.  A subspace estimator for fixed rank perturbations of large random matrices , 2011, J. Multivar. Anal..

[30]  M. Rudelson,et al.  Hanson-Wright inequality and sub-gaussian concentration , 2013 .

[31]  R. Couillet,et al.  Spectral analysis of the Gram matrix of mixture models , 2015, 1510.03463.

[32]  Romain Couillet,et al.  The random matrix regime of Maronna's M-estimator with elliptically distributed samples , 2013, J. Multivar. Anal..

[33]  F. Benaych-Georges,et al.  Lectures on the local semicircle law for Wigner matrices , 2016, 1601.04055.