- Research Article
- Open Access
The Convergence Rate for a -Functional in Learning Theory
Journal of Inequalities and Applications volume 2010, Article number: 249507 (2010)
It is known that in the field of learning theory based on reproducing kernel Hilbert spaces the upper bounds estimate for a -functional is needed. In the present paper, the upper bounds for the -functional on the unit sphere are estimated with spherical harmonics approximation. The results show that convergence rate of the -functional depends upon the smoothness of both the approximated function and the reproducing kernels.
It is known that the goal of learning theory is to approximate a function (or some function features) from data samples.
Let be a compact subset of -dimensional Euclidean spaces , . Then, learning theory is to find a function related the input to the output (see [1–3]). The function is determined by a probability distribution on where is the marginal distribution on and is the condition probability of for a given
Generally, the distribution is known only through a set of sample independently drawn according to . Given a sample , the regression problem based on Support Vector Machine (SVM) learning is to find a function such that is a good estimate of when a new input is provided. The binary classification problem based on SVM learning is to find a function which divides into two parts. Here is often induced by a real-valued function with the form of where if , otherwise, . The functions are often generated from the following Tikhonov regularization scheme (see, e.g., [4–9]) associated with a reproducing kernel Hilbert space (RKHS) (defined below) and a sample :
where is a positive constant called the regularization parameter and () called -norm SVM loss.
In addition, the Tikhonov regularization scheme involving offset (see, e.g., [4, 10, 11]) can be presented below with a similar way to (1.1)
We are in a position to define reproducing kernel Hilbert space. A function is called a Mercer kernel if it is continuous, symmetric, and positive semidefinite, that is, for any finite set of distinct points , the matrix is positive semidefinite.
The reproducing kernel Hilbert space (RKHS) (see ) associated with the Mercer kernel is defined to be the closure of the linear span of the set of functions with the inner product satisfying and the reproducing property
If , then . Denote as the space of continuous function on with the norm . Let Then the reproducing property tells that
It is easy to see that is a subset of We say that is a universal kernel if for any compact subset is dense in (see [13, Page 2652]).
Let be a given discrete set of finite points. Then, we may define an RKHS by the linear span of the set of functions . Then, it is easy to see that and for any there holds
Define and where the minimum is taken over all measurable functions. Then, to estimate the explicit learning rate, one needs to estimate the regularization errors (see, e.g., [4, 7, 9, 14])
The convergence rate of (1.5) is controlled by the -functional (see, e.g., )
and (1.6) is controlled by another -functional (see, e.g., )
We notice that, on one hand, the -functionals (1.7) and (1.8) are the modifications of the -functional of interpolation theory (see ) since the interpolation relation (1.4). On the other hand, they are different from the usual -functionals (see e.g., [16–30]) since the term However, they have some similar point. For example, if is a universal kernel, is dense in (see e.g., ). Moreover, some classical function spaces such as the polynomial spaces (see [2, 32]) and even some Sobolev spaces may be regarded as RKHS (see e.g., ).
In learning theory we often require and for some (see e.g., [1, 7, 14]). Many results on this topic have been achieved. With the weighted Durrmeyer operators [8, 9] showed the decay by taking to be the algebraic polynomials kernels on or on the simplex in .
However, in general case, the convergence of -functional (1.8) should also be considered since the offset often has influences on the solution of the learning algorithms (see e.g., [6, 11]). Hence, the purpose of this paper is twofold. One is to provide the convergence rates of (1.7) and (1.8) when is a general Mercer kernel on the unit sphere and The other is how to construct functions of the type of
to obtain the convergence rate of (1.8). The translation networks constructed in [34–37] have the form of (1.10) and the zonal networks constructed in [38, 39] have the form of (1.10) with . So the methods used by these references may be used here to estimate the convergence rates of (1.7) and (1.8) if one can bound the term
In the present paper, we shall give the convergence rate of (1.7) and (1.8) for a general kernel defined on the unit sphere and with being the usual Lebesgue measure on . If there is a distortion between and the convergence rate of (1.7)-(1.8) in the general case may be obtained according to the way used by [1, 8].
The rest of this paper is organized as follows. In Section 2, we shall restate some notations on spherical harmonics and present the main results. Some useful lemmas dealing with the approximation order for the de la Vallée means of the spherical harmonics, the Gauss integral formula, the Marcinkiewicz-Zygmund with respect to the scattered data obtained by G. Brown and F. Dai and a result on the zonal networks approximation provided by H. N. Mhaskar will be given in Section 3. A kind of weighted norm estimate for the Mercer kernel matrices on the unit sphere will be given in Lemma 3.8. Our main results are proved in the last section.
Throughout the paper, we shall write if there exists a constant such that . We write if and .
2. Notations and Results
To state the results of this paper, we need some notations and results on spherical harmonics.
For integers , , the class of all one variable algebraic polynomials of degree defined on is denoted by , the class of all spherical harmonics of degree will be denoted by , and the class of all spherical harmonics of degree will be denoted by . The dimension of is given by (see [40, Page 65])
and that of is One has the following well-known addition formula (see [41, Page 10, Theorem ]):
where is the degree- generalized Legendre polynomial. The Legendre polynomials are normalized so that and satisfy the orthogonality relations
Define and by taking to be the usual volume element of and the Jacobi weights functions , , , respectively. For any we have the following relation (see [42, Page 312]):
The orthogonal projections of a function on are defined by (see e.g., )
where denotes the inner product of and .
2.2. Main Results
Let satisfy and . Define
Then, by [44, Chapter 17] we know that is positive semidefinite on and the right of (2.6) is convergence absolutely and uniformly since . Therefore, is a Mercer kernel on By [13, Theorem ] we know that is a universal kernel on . We suppose that there is a constant depending only on such for any
Given a finite set , we denote by the cardinality of . For and we say that a finite subset is an -covering of if
where with being the geodesic distance between and .
Let be an integer, a sequence of real numbers. Define forward difference operators by , ,
We say a finite subset is a subset of interpolatory type if for any real numbers there is a such that , This kind of subsets may be found from [45, 46].
Let be the set of all sequence for which and the set of all sequence for which
Let be a real number, Then, we say if there is a function such that
We now give the results of this paper.
If there is a constant depending only on such that is a subset of interpolatory type and a -covering of satisfying with and being a given positive integer. is an integer. is a real number such that there is and , satisfies and . is the reproducing kernel space reproduced by and the kernel (2.6). . Then there is a constant depending only on and and a function with and a constant such that
The functions satisfying the conditions of Theorem 2.1 may be found in [39, Page 357].
Under the conditions of Theorem 2.1. If , then
Corollary 2.2 shows that the convergence rate of the -functional (1.8) is controlled by the smoothness of both the reproducing kernels and the approximated function .
If there is a constant depending only on such that is a subset of interpolatory type and a -covering of satisfying with and being a given positive integer. is the reproducing kernel space reproducing by and the kernel (2.6) with satisfying and Then, for and there holds
3. Some Lemmas
To prove Theorems 2.1 and 2.3, we need some lemmas. The first one is about the Gauss integral formula and Marcinkiewicz inequalities.
There exist constants depending only on such that for any positive integer and any -covering of satisfying , there exists a set of real numbers , such that
for any and for
where the constants of equivalence depending only on , , , and when is small. Here one employs the slight abuse of notation that
The second lemma we shall use is the Nikolskii inequality for the spherical harmonics.
Lemma 3.2 (see [38, 45, 49, 51, 52]).
If , , then one has the following Nikolskii inequality:
where the constant depends only on .
We now restate the general approximation frame of the Cesàro means and de la Vallée Poussin means provided by Dai and Ditzian (see ).
Let be a positive measure on . is a sequence of finite-dimensional spaces satisfying the following:
(II) is orthogonal to (in ) when
(III) is dense in for all .
(IV) is the collection of the constants.
The Cesàro means of is given by
for , where
and is an orthogonal base of in One sets,for a given , and if there exists such that
Let be defined as for and for and is a nonegative and nonincrease function. are the de la Vallée Poussin means defined as
Then, If for some , , then, and
Lemma 3.3 makes the following Lemma 3.4.
Let be the function defined as in Lemma 3.3. Define two kinds of operators, respectively, by
Then, for any and for any . Moreover,
where for one defines
By [54, Lemma ] we know for some . Hence, (3.9) holds by (3.7). By [19, Theorem ] we know for Hence, (3.10) holds by (3.7).
Let be a finite set. Then we call an M-Z quadrature measure of order if (3.1) and (3.2) hold for By this definition one knows the finite set in Lemma 3.1 is an M-Z quadrature measure of order .
Define an operator as
Then, we have the following results.
Lemma 3.5 (see ).
For a given integer let be an M-Z quadrature measure of order , , an integer, , , where satisfies which satisfies if and if . defined in Lemma 3.3 is a nonnegative and non-increasing function. Let satisfy . Then, for , , where consists of for which the derivative of order ; that is, , belongs to . Then, there is an operator such that
(i)(see [39, Proposition , (b)]). for
(ii)(see [39, Theorem ]). Moreover, if one adds an assumption that then, there are constants and such that
Lemma 3.6 (see e.g., [29, Page 230]).
Let . Then,
Following Lemma 3.7 deals with the orthogonality of the Legendre polynomials
For the generalized Legendre polynomials one has
It may be obtained by (2.2).
Let satisfy (2.7) for and . is a finite set satisfying the conditions of Theorem 2.1. Then, there is a constant depending only on such that
Define a matrix by , where with and Then,
By the Parseval equality we have
Let satisfy , . Then, by (3.1)
Hence, . On the other hand, since , , we have for any that
It follows for that
Define . Then, (3.24), (3.10), the Cauchy inequality, and the fact make
It follows that
Equation (3.2) thus holds.
4. Proof of the Main Results
We now show Theorems 2.1 and 2.3, respectively.
Proof of Theorem 2.1.
Lemma in  gave the following results.
Let,, be an integer, and a sequence of real numbers such . Then, there exists such that ,
Since and we have a such that Hence, and
and for there holds for that
It follows for that
On the other hand, since
where for , we have by (4.3)
Hence, above equation and (3.1)-(3.2) makes
where , Define
Then, we know and by (3.9)
It follows by (3.9) that
On the other hand, by the definition of and (3.14) we have for that
where denotes the operator of Lemma 3.5 for Hence,
Equation (3.2) and the definition of make
The Hölder inequality, the of Lemma 3.5, and the fact that make . Therefore,
Equations (3.2), (3.17), (3.16), and the Cauchy inequality make
Let be the Gamma function. Then, it is well known that Therefore,
Equations (4.14) and (4.4) make
Since , we have (2.11) by (4.20). Equation (2.12) follows by (4.3), (4.4), and (3.19).
Proof of Corollary 2.2.
By (2.11)-(2.12) one has
Proof of Theorem 2.3.
Take the place of in Lemma 3.5 with denote still by the operator in Lemma 3.5 with and
then, and by (3.15) In this case,
Since is a spherical harmonics of order , we know by of Lemma 3.5 that are also spherical harmonics of order Then, (3.2), of Lemma 3.5, (3.3), and (3.16) make
Hence, (3.19) and above equation make . Equation (2.14) follows by (3.15).
Cucker F, Smale S: On the mathematical foundations of learning. Bulletin of the American Mathematical Society 2002, 39(1):1–49.
Cucker F, Zhou D-X: Learning Theory: An Approximation Theory Viewpoint, Cambridge Monographs on Applied and Computational Mathematics. Cambridge University Press, Cambridge, Mass, USA; 2007:xii+224.
Vapnik VN: Statistical Learning Theory, Adaptive and Learning Systems for Signal Processing, Communications, and Control. John Wiley & Sons, New York, NY, USA; 1998:xxvi+736.
Chen DR, Wu Q, Ying YM, Zhou DX: Support vector machine soft margin classifier: error analysis. Journal of Machine Learning and Research 2004, 5: 1143–1175.
Evgeniou T, Pontil M, Poggio T: Regularization networks and support vector machines. Advances in Computational Mathematics 2000, 13(1):1–50. 10.1023/A:1018946025316
Li Y, Liu Y, Zhu J: Quantile regression in reproducing kernel Hilbert spaces. Journal of the American Statistical Association 2007, 102(477):255–268. 10.1198/016214506000000979
Tong H, Chen D-R, Peng L: Analysis of support vector machines regression. Foundations of Computational Mathematics 2009, 9(2):243–257. 10.1007/s10208-008-9026-0
Tong H, Chen D-R, Peng L: Learning rates for regularized classifiers using multivariate polynomial kernels. Journal of Complexity 2008, 24(5–6):619–631. 10.1016/j.jco.2008.05.008
Zhou D-X, Jetter K: Approximation with polynomial kernels and SVM classifiers. Advances in Computational Mathematics 2006, 25(1–3):323–344.
Chen D, Xiang D-H: The consistency of multicategory support vector machines. Advances in Computational Mathematics 2006, 24(1–4):155–169.
De Vito E, Rosasco L, Caponnetto A, Piana M, Verri A: Some properties of regularized kernel methods. Journal of Machine Learning Research 2004, 5: 1363–1390.
Aronszajn N: Theory of reproducing kernels. Transactions of the American Mathematical Society 1950, 68: 337–404. 10.1090/S0002-9947-1950-0051437-7
Micchelli CA, Xu Y, Zhang H: Universal kernels. Journal of Machine Learning Research 2006, 7: 2651–2667.
Wu Q, Ying Y, Zhou D-X: Multi-kernel regularized classifiers. Journal of Complexity 2007, 23(1):108–134. 10.1016/j.jco.2006.06.007
Bergh J, Löfström J: Interpolation Spaces. Springer, New York, NY, USA; 1976.
Berens H, Lorentz GG: Inverse theorems for Bernstein polynomials. Indiana University Mathematics Journal 1972, 21(8):693–708. 10.1512/iumj.1972.21.21054
Berens H, Li LQ: The Peetre -moduli and best approximation on the sphere. Acta Mathematica Sinica 1995, 38(5):589–599.
Berens H, Xu Y: -moduli, moduli of smoothness, and Bernstein polynomials on a simplex. Indagationes Mathematicae 1991, 2(4):411–421. 10.1016/0019-3577(91)90027-5
Chen W, Ditzian Z: Best approximation and -functionals. Acta Mathematica Hungarica 1997, 75(3):165–208. 10.1023/A:1006543020828
Chen W, Ditzian Z: Best polynomial and Durrmeyer approximation in . Indagationes Mathematicae 1991, 2(4):437–452. 10.1016/0019-3577(91)90029-7
Dai F, Ditzian Z: Jackson inequality for Banach spaces on the sphere. Acta Mathematica Hungarica 2008, 118(1–2):171–195. 10.1007/s10474-007-6206-3
Ditzian Z, Zhou X: Optimal approximation class for multivariate Bernstein operators. Pacific Journal of Mathematics 1993, 158(1):93–120.
Ditzian Z, Runovskii K: Averages and -functionals related to the Laplacian. Journal of Approximation Theory 1999, 97(1):113–139. 10.1006/jath.1997.3262
Ditzian Z: A measure of smoothness related to the Laplacian. Transactions of the American Mathematical Society 1991, 326(1):407–422. 10.2307/2001870
Ditzian Z, Totik V: Moduli of Smoothness, Springer Series in Computational Mathematics. Volume 9. Springer, New York, NY, USA; 1987:x+227.
Ditzian Z: Approximation on Banach spaces of functions on the sphere. Journal of Approximation Theory 2006, 140(1):31–45. 10.1016/j.jat.2005.11.013
Ditzian Z: Fractional derivatives and best approximation. Acta Mathematica Hungarica 1998, 81(4):323–348. 10.1023/A:1006554907440
Schumaker LL: Spline Functions: Basic Theory. John Wiley & Sons, New York, NY, USA; 1981:xiv+553. Pure and Applied Mathematics Pure and Applied Mathematics
Wang KY, Li LQ: Harmonic Analysis and Approximation on the Unit Sphere. Science Press, Beijing, China; 2000.
Xu Y: Approximation by means of -harmonic polynomials on the unit sphere. Advances in Computational Mathematics 2004, 21(1–2):37–58.
Smale S, Zhou D-X: Estimating the approximation error in learning theory. Analysis and Applications 2003, 1(1):17–41. 10.1142/S0219530503000089
Sheng B: Estimates of the norm of the Mercer kernel matrices with discrete orthogonal transforms. Acta Mathematica Hungarica 2009, 122(4):339–355. 10.1007/s10474-008-8037-2
Loustau S: Aggregation of SVM classifiers using Sobolev spaces. Journal of Machine Learning Research 2008, 9: 1559–1582.
Mhaskar HN, Micchelli CA: Degree of approximation by neural and translation networks with a single hidden layer. Advances in Applied Mathematics 1995, 16(2):151–183. 10.1006/aama.1995.1008
Sheng BH: Approximation of periodic functions by spherical translation networks. Acta Mathematica Sinica. Chinese Series 2007, 50(1):55–62.
Sheng B: On the degree of approximation by spherical translations. Acta Mathematicae Applicatae Sinica. English Series 2006, 22(4):671–680. 10.1007/s10255-006-0341-4
Sheng B, Wang J, Zhou S: A way of constructing spherical zonal translation network operators with linear bounded operators. Taiwanese Journal of Mathematics 2008, 12(1):77–92.
Mhaskar HN, Narcowich FJ, Ward JD: Approximation properties of zonal function networks using scattered data on the sphere. Advances in Computational Mathematics 1999, 11(2–3):121–137.
Mhaskar HN: Weighted quadrature formulas and approximation by zonal function networks on the sphere. Journal of Complexity 2006, 22(3):348–370. 10.1016/j.jco.2005.10.003
Groemer H: Geometric Applications of Fourier Series and Spherical Harmonics, Encyclopedia of Mathematics and Its Applications. Volume 61. Cambridge University Press, Cambridge, Mass, USA; 1996:xii+329.
Müller C: Spherical Harmonics, Lecture Notes in Mathematics. Volume 17. Springer, Berlin, Germany; 1966:iv+45.
Lu SZ, Wang KY: Bochner-Riesz Means. Beijing Normal University Press, Beijing, China; 1988.
Wang Y, Cao F: The direct and converse inequalities for jackson-type operators on spherical cap. Journal of Inequalities and Applications 2009, 2009:-16.
Wendland H: Scattered Data Approximation, Cambridge Monographs on Applied and Computational Mathematics. Volume 17. Cambridge University Press, Cambridge, Mass, USA; 2005:x+336.
Mhaskar HN, Narcowich FJ, Sivakumar N, Ward JD: Approximation with interpolatory constraints. Proceedings of the American Mathematical Society 2002, 130(5):1355–1364. 10.1090/S0002-9939-01-06240-2
Narcowich FJ, Ward JD: Scattered data interpolation on spheres: error estimates and locally supported basis functions. SIAM Journal on Mathematical Analysis 2002, 33(6):1393–1410. 10.1137/S0036141001395054
Brown G, Dai F: Approximation of smooth functions on compact two-point homogeneous spaces. Journal of Functional Analysis 2005, 220(2):401–423. 10.1016/j.jfa.2004.10.005
Brown G, Feng D, Sheng SY: Kolmogorov width of classes of smooth functions on the sphere . Journal of Complexity 2002, 18(4):1001–1023. 10.1006/jcom.2002.0656
Dai F: Multivariate polynomial inequalities with respect to doubling weights and weights. Journal of Functional Analysis 2006, 235(1):137–170. 10.1016/j.jfa.2005.09.009
Mhaskar HN, Narcowich FJ, Ward JD: Spherical Marcinkiewicz-Zygmund inequalities and positive quadrature. Mathematics of Computation 2001, 70(235):1113–1130.
Belinsky E, Dai F, Ditzian Z: Multivariate approximating averages. Journal of Approximation Theory 2003, 125(1):85–105. 10.1016/j.jat.2003.09.005
Kamzolov AI: Approximation of functions on the sphere . Serdica 1984, 10(1):3–10.
Dai F, Ditzian Z: Cesàro summability and Marchaud inequality. Constructive Approximation 2007, 25(1):73–88. 10.1007/s00365-005-0623-8
Dai F: Some equivalence theorems with -functionals. Journal of Approximation Theory 2003, 121(1):143–157. 10.1016/S0021-9045(02)00059-X
This work is supported by the National NSF (10871226) of China. The authors thank the reviewers for giving very valuable suggestions.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 2.0 International License (https://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
About this article
Cite this article
Sheng, BH., Xiang, DH. The Convergence Rate for a -Functional in Learning Theory. J Inequal Appl 2010, 249507 (2010). https://doi.org/10.1155/2010/249507
- Spherical Harmonic
- Tikhonov Regularization
- Cauchy Inequality
- Jacobi Weight
- Mercer Kernel