Identifiability in penalized function-on-function regression models
From MaRDI portal
Publication:259185
DOI10.1214/16-EJS1123zbMATH Open1332.62249arXiv1506.03627OpenAlexW1927268375MaRDI QIDQ259185FDOQ259185
Authors: Fabian Scheipl, Sonja Greven 
Publication date: 11 March 2016
Published in: Electronic Journal of Statistics (Search for Journal in Brave)
Abstract: Regression models with functional responses and covariates constitute a powerful and increasingly important model class. However, regression with functional data poses well known and challenging problems of non-identifiability. This non-identifiability can manifest itself in arbitrarily large errors for coefficient surface estimates despite accurate predictions of the responses, thus invalidating substantial interpretations of the fitted models. We offer an accessible rephrasing of these identifiability issues in realistic applications of penalized linear function-on-function-regression and delimit the set of circumstances under which they are likely to occur in practice. Specifically, non-identifiability that persists under smoothness assumptions on the coefficient surface can occur if the functional covariate's empirical covariance has a kernel which overlaps that of the roughness penalty of the spline estimator. Extensive simulation studies validate the theoretical insights, explore the extent of the problem and allow us to evaluate their practical consequences under varying assumptions about the data generating processes. A case study illustrates the practical significance of the problem. Based on theoretical considerations and our empirical evaluation, we provide immediately applicable diagnostics for lack of identifiability and give recommendations for avoiding estimation artifacts in practice.
Full work available at URL: https://arxiv.org/abs/1506.03627
Recommendations
- Penalized function-on-function regression
- scientific article; zbMATH DE number 962363
- Penalized likelihood functional regression
- Penalized solutions to functional regression problems
- Penalized Estimating Functions and Variable Selection in Semiparametric Regression Models
- Penalized spline estimation for functional coefficient regression models
- Robust penalized estimators for functional linear regression
- Estimation in generalized linear models for functional data via penalized likelihood
- Identification and estimation by penalization in nonparametric instrumental regression
Diagnostics, and linear inference and regression (62J20) Ridge regression; shrinkage estimators (Lasso) (62J07)
Cites Work
- Functional data analysis.
- Low‐Rank Scale‐Invariant Tensor Product Smooths for Generalized Additive Mixed Models
- Functional linear model
- Penalized function-on-function regression
- Functional Principal Component Regression and Functional Partial Least Squares
- Title not available (Why is that?)
- Practical variable selection for generalized additive models
- Functional linear regression analysis for longitudinal data
- Generalized Linear Models with Functional Predictors
- Title not available (Why is that?)
- A distance measure between cointegration spaces
- Generalized multilevel functional regression
- Functional quadratic regression
- Title not available (Why is that?)
- Functional Adaptive Model Estimation
- Varying-coefficient functional linear regression
Cited In (22)
- Comments on: ``Shape-based functional data analysis
- Spike-and-slab priors for function selection in structured additive regression models
- Smooth LASSO estimator for the function-on-function linear regression model
- Envelope Model for Function-on-Function Linear Regression
- A partial least squares approach for function-on-function interaction regression
- Robust depth-based estimation of the functional autoregressive model
- Statistical inference for the slope parameter in functional linear regression
- Adaptive smoothing spline estimator for the function-on-function linear regression model
- Optimal Penalized Function-on-Function Regression Under a Reproducing Kernel Hilbert Space Framework
- Advances in estimation and inference for historical functional linear models
- Bayesian wavelet-packet historical functional linear models
- Statistical inference for function-on-function linear regression
- Comparison and contrast of two general functional regression modelling frameworks
- Latent group detection in functional partially linear regression models
- Functional Regression for Densely Observed Data With Novel Regularization
- Bayesian function-on-function regression for multilevel functional data
- Boosting flexible functional regression models with a high number of functional historical effects
- A general framework for functional regression modelling
- On the estimation of functional random effects
- Estimation and inference for functional linear regression models with partially varying regression coefficients
- Variable Selection for Multiple Function-on-Function Linear Regression
- Sparse functional linear models via calibrated concave-convex procedure
Uses Software
This page was built for publication: Identifiability in penalized function-on-function regression models
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q259185)
