论文信息 - The Stability of Kernel Principal Components Analysis and its Relation to the Process Eigenspectrum

The Stability of Kernel Principal Components Analysis and its Relation to the Process Eigenspectrum

In this paper we analyze the relationships between the eigenvalues of the m x m Gram matrix K for a kernel k (·, ·) corresponding to a sample x1,...,xm drawn from a density p(x) and the eigenvalues of the corresponding continuous eigenproblem. We bound the differences between the two spectra and provide a performance bound on kernel PCA.

John Shawe-Taylor | Christopher K. I. Williams | J. Shawe-Taylor

[1] T. W. Anderson. ASYMPTOTIC THEORY FOR PRINCIPAL COMPONENT ANALYSIS , 1963 .

[2] R. Taylor,et al. The Numerical Treatment of Integral Equations , 1978 .

[3] Christopher K. I. Williams,et al. Gaussian regression and optimal finite dimensional linear models , 1997 .

[4] Bernhard Schölkopf,et al. Nonlinear Component Analysis as a Kernel Eigenvalue Problem , 1998, Neural Computation.

[5] Christopher K. I. Williams,et al. The Effect of the Input Density Distribution on Kernel-based Classifiers , 2000, ICML.

[6] V. Koltchinskii,et al. Random matrix approximation of spectra of integral operators , 2000 .

[7] Nello Cristianini,et al. On the Concentration of Spectral Properties , 2001, NIPS.