Summary
Recent literature on functional estimation has shown the importance of kernels with vanishing moments although no general framework was given to build kernels of increasing order apart from some specific methods based on moment relationships. The purpose of the present paper is to develop such a framework and to show how to build higher order kernels with nice properties and to solve optimization problems about kernels. The proofs given here, unlike standard variational arguments, explain why some hierarchies of kernels do have optimality properties. Applications are given to functional estimation in a general context. In the last section special attention is paid to density estimates based on kernels of order (m, r), i.e., kernels of orderr for estimation of derivatives of orderm. Convergence theorems are easily derived from interpretation by means of projections inL 2 spaces.
Article PDF
Similar content being viewed by others
References
Berlinet, A.: Reproducing kernels and finite order kernels. In: Roussas, G. (ed.) Nonparametric functional estimation and related topics, pp. 3–18. London New York: Kluwer 1990
Berlinet, A., Devroye, L.: Estimation d'une densité: un point sur la méthode du noyau. Stat. Anal. Données,14 (no 1), 1–32 (1989)
Brézinski, C.: Padé-type approximation and general orthogonal polynomials. Basel: Birkhäuser 1980
Deheuvels, P.: Estimation non-paramétrique de la densité par histogrammes généralisés. Rev. Stat. Appl.25, 5–42 (1977)
Devroye, L.: The double kernel method in density estimation. Ann. Inst. Henri Poincaré25 (no 4), 553–580 (1989)
Epanechnikov, V.A.: Nonparametric estimation of a multidimensional probability density. Theory Probab. Appl.14, 153–158 (1969)
Freud, G.: On polynomial approximation with the weight exp (−x 2k/2). Acta Math. Acad. Sci. Hung.24, 363–371 (1973)
Gasser, T., Müller, H.G.: Kernel estimation of regression function. In: Gasser, T., Rosenblatt, M. (eds.) Smoothing techniques for curve estimation. (Lect. Notes Math., vol. 757, pp. 23–68) Berlin Heidelberg New York: Springer 1979a
Gasser, T., Müller, H.G.: Optimal convergence properties of kernel estimates of derivatives of a density function. In: Gasser, T., Rosenblatt, M. (eds.) Smoothing techniques for curve estimation. (Lect. Notes Math., vol. 757, pp. 144–154) Berlin Heidelberg New York: Springer 1979b
Gasser, T., Müller, H.G., Mammitzsch, V.: Kernels for nonparametric curve estimation. J. R. Stat. Soc., Ser. B47, 238–252 (1985)
Granovsky, B.L., Müller, H.G.: On the optimality of a class of polynomial kernel functions. Stat. Decis.7, 301–312 (1989)
Granovsky, B.L., Müller, H.G.: Optimizing kernel methods: a unifying variational principle. Int. Stat. Rev.59 (3), 373–388 (1991)
Hall, P., Marron, J.S.: Choice of kernel order in density estimation. Ann. Stat.16, 161–173 (1988)
Jones, M.C.: Changing kernels' orders. (Preprint 1990)
Mc Diarmid, C.: On the method of bounded differences. In: Surveys in combinatorics. (Lond. Math. Soc. Lect. Notes Ser., vol. 141, pp. 148–188) Cambridge: Cambridge University Press 1989
Mammitzsch, V.: A note on kernels of orderv, k. In: Mande, P., Huškovich, M. (eds.) Proceedings of the Fourth Prague Symposium on Asymptotic Statistics, pp. 411–412. Prague: Charles University 1989
Müller, H.G.: Smooth optimum kernel estimators of densities, regression curves and modes. Ann. Stat.12, 766–774 (1984)
Müller, H.G.: Weighted local regression and kernel methods for nonparametric curve fitting. J. Am. Stat. Assoc.82, 231–238 (1987)
Müller, H.G.: On the construction of boundary kernels. University of California at Davis (Preprint 1991)
Nevai, P.: Some properties of orthogonal polynomials corresponding to the weight (1+x 2k)αexp(−x 2k) and their application in approximation theory. Sov. Math. Dokl.14, 1116–1119 (1973a)
Nevai, P.: Orthogonal polynomials on the real line associated with the weight |x|αexp(−|x|β), I. Acta Math. Acad. Sci. Hung.24, 335–342 (1973b)
Nevai, P.: Orthogonal polynomials. Mem. Am. Math. Soc.219 (1979)
Schucany, W.R.: On nonparametric regression with higher-order kernels. J. Stat. Plann. Inference23, 145–151 (1989)
Schucany, W.R.: Sommers, J.P.: Improvement of kernel type density estimators. J. Am. Stat. Assoc.72, 420–423 (1977)
Silverman, B.W.: Density estimation for statistics and data analysis. London: Chapman and Hall 1986
Singh, R.S.: Mean squared errors of estimates of a density and its derivatives. Biometrika66(1), 177–180 (1979)
Stuetzle, W., Mittal, Y.: Some comments on the asymptotic behavior of robust smoothers. In: Gasser, T., Rosenblatt, M. (eds.) Smoothing techniques for curve estimation. (Lect. Notes Math., vol. 757, pp. 191–195) Berlin Heidelberg New York: Springer 1979
Wand, M., Schucany, W.R.: Gaussian-based kernels. Can. J. Stat.18, 197–204 (1990)