Additive regression for predictors of various natures and possibly incomplete Hilbertian responses

Abstract: In this paper we consider a fully nonparametric additive regression model for responses and predictors of various natures. This includes the case of Hilbertian and incomplete (like censored or missing) responses, and continuous, nominal discrete and ordinal discrete predictors. We propose a backfitting technique that estimates this additive model, and establish the existence of the estimator and the convergence of the associated backfitting algorithm under minimal conditions. We also develop a general asymptotic theory for the estimator such as the rates of convergence and asymptotic distribution. We verify the practical performance of the proposed estimator in a simulation study. We also apply the method to various real data sets, including those for a density-valued response regressed on a mixture of continuous and nominal discrete predictors, for a compositional response regressed on a mixture of continuous and ordinal discrete predictors, and for a censored scalar response regressed on a mixture of continuous and nominal discrete predictors.

[1]  E. Mammen,et al.  Backfitting and smooth backfitting for additive quantile models , 2010, 1011.2592.

[2]  M. Di Marzio,et al.  Non-parametric regression for compositional data , 2015 .

[3]  Alessandra Menafoglio,et al.  Compositional regression with functional response , 2018, Comput. Stat. Data Anal..

[4]  James O. Ramsay,et al.  Applied Functional Data Analysis: Methods and Case Studies , 2002 .

[5]  K. Mukherjee,et al.  Central limit theorems revisited , 2000 .

[6]  I. Keilegom,et al.  Varying coefficient models having different smoothing variables with randomly censored data , 2014 .

[7]  Enno Mammen,et al.  Semi-parametric regression: Efficiency gains from modeling the nonparametric part , 2011, 1104.4410.

[8]  Enno Mammen,et al.  The Existence and Asymptotic Properties of a Backfitting Projection Algorithm Under Weak Conditions , 1999 .

[9]  E. Mammen,et al.  A simple smooth backfitting method for additive models , 2006, math/0702657.

[10]  J. V. Ryzin,et al.  Regression Analysis with Randomly Right-Censored Data , 1981 .

[11]  Enno Mammen,et al.  Flexible generalized varying coefficient regression models , 2012, 1210.4711.

[12]  H. Muller,et al.  Functional data analysis for density functions by transformation to a Hilbert space , 2016, 1601.02869.

[13]  Robert Kohn,et al.  Convergence of the backfitting algorithm for additive models , 1994 .

[14]  Denis Bosq,et al.  Linear Processes in Function Spaces , 2000 .

[15]  Hans-Georg Müller,et al.  Functional Data Analysis , 2016 .

[16]  Enno Mammen,et al.  Projection-type estimation for varying coefficient regression models , 2012, 1203.0403.

[17]  Jeong Min Jeon,et al.  Additive regression with Hilbertian responses , 2020, The Annals of Statistics.

[18]  J. Aitchison,et al.  Multivariate binary discrimination by the kernel method , 1976 .

[19]  V. Pawlowsky-Glahn,et al.  Compositional data analysis : theory and applications , 2011 .

[20]  Michail Tsagris,et al.  Regression analysis with compositional data containing zero values , 2015, 1508.01913.

[21]  A. Buja.  What Criterion for a Power Algorithm , 1996 .

[22]  A Note on The Backfitting Estimation of Additive Models , 2009, 0903.3470.

[23]  David Ruppert,et al.  Fitting a Bivariate Additive Model by Local Polynomial Regression , 1997 .

[24]  K. Singh,et al.  The product-limit estimator and the bootstrap: Some asymptotic representations , 1986 .

[25]  Kim C. Border,et al.  Infinite Dimensional Analysis: A Hitchhiker’s Guide , 1994 .

[26]  E. J. Beltrami,et al.  On Infinite-Dimensional Convex Programs , 1967, J. Comput. Syst. Sci..

[27]  Enno Mammen,et al.  Smooth backfitting in generalized additive models , 2008, 0803.1922.

[28]  P. Speckman.  Kernel smoothing in partial linear models , 1988 .

[29]  Heng Lian,et al.  Convergence of functional k-nearest neighbor regression estimate with functional responses , 2011, Electronic Journal of Statistics.

[30]  Peter Filzmoser,et al.  Simplicial principal component analysis for density functions in Bayes spaces , 2016, Comput. Stat. Data Anal..

[31]  I. Keilegom,et al.  Estimation of a semiparametric transformation model , 2008, 0804.0719.

[32]  Ingrid Van Keilegom,et al.  Regression when both response and predictor are functions , 2012, J. Multivar. Anal..

[33]  Jean D. Opsomer,et al.  Asymptotic Properties of Backfitting Estimators , 2000 .

[34]  Elias Masry,et al.  MULTIVARIATE LOCAL POLYNOMIAL REGRESSION FOR TIME SERIES:UNIFORM STRONG CONSISTENCY AND RATES , 1996 .

[35]  H. Muller,et al.  Wasserstein covariance for multiple random densities , 2018, Biometrika.

[36]  Stanley R. Johnson,et al.  Varying Coefficient Models , 1984 .

[37]  O. Linton,et al.  A kernel method of estimating structured nonparametric regression based on marginal integration , 1995 .

[38]  V. Pawlowsky-Glahn,et al.  Compositional data: the sample space and its structure , 2019, TEST.

[39]  Jia-gang Wang,et al.  A Note on the Uniform Consistency of the Kaplan-Meier Estimator , 1987 .

[40]  H. Müller,et al.  Additive Functional Regression for Densities as Responses , 2020, Journal of the American Statistical Association.