MULTIRESOLUTION ANALYSES AND WAVELETS FOR DENSITY ESTIMATION

Abstract. We use the concept of multiresolution analysis and orthonormal basisof wavelet functions to construct an estimator for an unknown density in a givenSobolev’s space. We show consistency with the rate of convergence equal to theoptimal minimax rate of convergence. We also discuss some of the features of thisnew method of estimation on a numerical example. 1. IntroductionLet X 1 ,X 2 ,··· ,X n be independent identically distributed random variables, thecommon distribution having the density f with respect to the Lebesque measure ofIR d . Many methods for the estimation of the unknown density f have been proposedand used over the years. They include the kernel methods (Rozenblatt [14], Parzen[13]. Watson and Leabetter [22], the orthogonal series estimators used by Cencov [3],Kronmal and Tarter [10], Schwartz [16], Walter ([19],[20]), Walter and Blum [21], theinterpolation methods considered by Wahba ([17], [18]). We refer the reader to theextensive bibliographies of the textbooks [5] and [15] on density estimation.The kernel method of density estimation is presumably the most popular of all ofthem. Its local character is certainly its main attraction. The method of orthogonal

[1]  L. Devroye,et al.  Nonparametric density estimation : the L[1] view , 1987 .

[2]  R. Kronmal,et al.  The Estimation of Probability Densities and Cumulatives by Fourier Series Methods , 1968 .

[3]  G. Wahba Optimal Convergence Properties of Variable Knot, Kernel, and Orthogonal Series Methods for Density Estimation. , 1975 .

[4]  E. Parzen On Estimation of a Probability Density Function and Mode , 1962 .

[5]  G. Walter Approximation of the delta function by wavelets , 1992 .

[6]  J. Bretagnolle,et al.  Estimation des densités: risque minimax , 1978 .

[7]  S. Mallat Multiresolution approximations and wavelet orthonormal bases of L^2(R) , 1989 .

[8]  M. Rosenblatt Remarks on Some Nonparametric Estimates of a Density Function , 1956 .

[9]  G. Walter,et al.  Probability Density Estimation Using Delta Sequences , 1979 .

[10]  G. Wahba,et al.  The computation of generalized cross-validation functions through householder tridiagonalization with applications to the fitting of interaction spline models , 1989 .

[11]  P. J. Green,et al.  Density Estimation for Statistics and Data Analysis , 1987 .

[12]  M. R. Leadbetter,et al.  On the Estimation of the Probability Density, I , 1963 .

[13]  I. Daubechies Orthonormal bases of compactly supported wavelets , 1988 .

[14]  S. Schwartz Estimation of Probability Density by an Orthogonal Series , 1967 .

[15]  G. Walter Properties of Hermite Series Estimation of Probability Density , 1977 .

[16]  C. J. Stone,et al.  A study of logspline density estimation , 1991 .

[17]  G. Wahba Optimal Smoothing of Density Estimates , 1977 .