Behavior of a functional in learning theory

Abstract

Let H be a Hilbert space, A ∈ L(H), y ∈ $\overline {R(A)} $$, and y ∉ R(A). We study the behavior of the distance square between y and A(B_τ), defined as a functional F(τ), as the radius τ of the ball B_τ of H tends to ∞. This problem is important in estimating the approximation error in learning theory. Our main result is to estimate the asymptotic behavior of F(τ) without the compactness assumption on the operator A. We also consider the Peetre K-functional and its convergence rates.

Keywords

Learning theory / approximation error / Peetre K-functional / reproducing kernel Hilbert space / regression learning problem

References

[1] Cucker F., Smale S. On the mathematical foundations of learning. Bull Amer Soc, 2001, 39: 1-49.

[2] Smale S., Zhou D. X. Estimating the approximation error in learning theory. Anal Appl, 2003, 1: 17-41.

[3] Vapnik V. Statistical Learning Theory, 1998, New York: Wiley.

[4] Micchelli C. A., Pinkus A. Variational problems arising from balancing several error criteria. Rendiconti di Matematica, Serie VII, 1994, 14: 37-86.

[5] Chen D. R., Xiang D. H. The consistency of multicategory support vector machines. Adv Comput Math, 2006, 24: 155-169.

[6] Sun Z. H., Sun Y. X. Least square generalized support vector machines for regression. Acta Math, 2004, 24: 94-97.

[7] Wu Z. M., Schaback R. Local error estimates for radial basis function interpolation of scattered data. IMA J Numer Anal, 1993, 13: 13-27.

[8] Smale S., Zhou D. X. Shannon sampling II. Connections to learning theory. Appl Comput Harmonic Anal, 2005, 19: 285-302.

[9] Micchelli C. A., Pontil M. Learning the kernel function via regularization. Journal of Machine Learning Research, 2005, 6: 1099-1125.