Accurate Error Bounds for the Eigenvalues of the Kernel Matrix
نویسنده
چکیده
The eigenvalues of the kernel matrix play an important role in a number of kernel methods, in particular, in kernel principal component analysis. It is well known that the eigenvalues of the kernel matrix converge as the number of samples tends to infinity. We derive probabilistic finite sample size bounds on the approximation error of individual eigenvalues which have the important property that the bounds scale with the eigenvalue under consideration, reflecting the actual behavior of the approximation errors as predicted by asymptotic results and observed in numerical simulations. Such scaling bounds have so far only been known for tail sums of eigenvalues. Asymptotically, the bounds presented here have a slower than stochastic rate, but the number of sample points necessary to make this disadvantage noticeable is often unrealistically large. Therefore, under practical conditions, and for all but the largest few eigenvalues, the bounds presented here form a significant improvement over existing non-scaling bounds.
منابع مشابه
Eigenvalues Ratio for Kernel Selection of Kernel Methods
The selection of kernel function which determines the mapping between the input space and the feature space is of crucial importance to kernel methods. Existing kernel selection approaches commonly use some measures of generalization error, which are usually difficult to estimate and have slow convergence rates. In this paper, we propose a novel measure, called eigenvalues ratio (ER), of the ti...
متن کاملSome remarks on the sum of the inverse values of the normalized signless Laplacian eigenvalues of graphs
Let G=(V,E), $V={v_1,v_2,ldots,v_n}$, be a simple connected graph with $%n$ vertices, $m$ edges and a sequence of vertex degrees $d_1geqd_2geqcdotsgeq d_n>0$, $d_i=d(v_i)$. Let ${A}=(a_{ij})_{ntimes n}$ and ${%D}=mathrm{diag }(d_1,d_2,ldots , d_n)$ be the adjacency and the diagonaldegree matrix of $G$, respectively. Denote by ${mathcal{L}^+}(G)={D}^{-1/2}(D+A) {D}^{-1/2}$ the normalized signles...
متن کاملAsymptotic error bounds for kernel-based Nystrm low-rank approximation matrices
• Many kernel-based learning algorithms have the computational load. • The Nyström low-rank approximation is designed for reducing the computation. • We propose the spectrum decomposition condition with a theoretical justification. • Asymptotic error bounds on eigenvalues and eigenvectors are derived. • Numerical experiments are provided for covariance kernel and Wishart matrix. AMS subject cla...
متن کاملSharp Bounds on the PI Spectral Radius
In this paper some upper and lower bounds for the greatest eigenvalues of the PI and vertex PI matrices of a graph G are obtained. Those graphs for which these bounds are best possible are characterized.
متن کاملAPPLICATION OF THE RANDOM MATRIX THEORY ON THE CROSS-CORRELATION OF STOCK PRICES
The analysis of cross-correlations is extensively applied for understanding of interconnections in stock markets. Variety of methods are used in order to search stock cross-correlations including the Random Matrix Theory (RMT), the Principal Component Analysis (PCA) and the Hierachical Structures. In this work, we analyze cross-crrelations between price fluctuations of 20 company stocks...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
- Journal of Machine Learning Research
دوره 7 شماره
صفحات -
تاریخ انتشار 2006