Functional additive regression

Abstract

We suggest a new method, called Functional Additive Regression, or FAR, for efficiently performing high-dimensional functional regression. FAR extends the usual linear regression model involving a functional predictor, $X(t)$, and a scalar response, $Y$, in two key respects. First, FAR uses a penalized least squares optimization approach to efficiently deal with high-dimensional problems involving a large number of functional predictors. Second, FAR extends beyond the standard linear regression setting to fit general nonlinear additive models. We demonstrate that FAR can be implemented with a wide range of penalty functions using a highly efficient coordinate descent algorithm. Theoretical results are developed which provide motivation for the FAR optimization criterion. Finally, we show through simulations and two real data sets that FAR can significantly outperform competing methods.

Published at http://dx.doi.org/10.1214/15-AOS1346 in the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)

References in corpus (7)

• On the "degrees of freedom" of the lasso
• One-step sparse estimates in nonconcave penalized likelihood models
• Variable selection in nonparametric additive models
• A unified approach to model selection and sparse recovery using regularized least squares
• Single and multiple index functional regression models with nonparametric link
• Functional additive regression
• Functional response additive model estimation with online virtual stock markets

Cited by in corpus (23)

• Functional additive regression
• Grouped variable importance with random forests and application to multiple functional data analysis
• Non-linear Functional Modeling using Neural Networks
• Variable selection in functional regression models: a review
• Functional response additive model estimation with online virtual stock markets
• Modern Non-Linear Function-on-Function Regression
• An autocovariance-based learning framework for high-dimensional functional time series
• Conditional Independence Testing in Hilbert Spaces with Applications to Functional Data Analysis
• Modern Multiple Imputation with Functional Data
• Shape-constrained Estimation in Functional Regression with Bernstein Polynomials
• FAStEN: An Efficient Adaptive Method for Feature Selection and Estimation in High-Dimensional Functional Regressions
• Functional Linear Regression with Mixed Predictors
• Finite Sample Theory for High-Dimensional Functional/Scalar Time Series with Applications
• Efficient Bayesian hierarchical functional data analysis with basis function approximations using Gaussian-Wishart processes
• Classification of pediatric pneumonia using chest X-rays by functional regression
• The Function-on-Scalar LASSO with Applications to Longitudinal GWAS
• Varying-coefficient functional additive models
• Dimension Independent Generalization Error by Stochastic Gradient Descent
• A General Theory for Large-Scale Curve Time Series via Functional Stability Measure
• Aggregating multiple types of complex data in stock market prediction: A model-independent framework
• Nonlinear Mixed-effects Scalar-on-function Models and Variable Selection for Kinematic Upper Limb Movement Data
• Tree-based boosting with functional data
• Estimation of a functional single index model with dependent errors and unknown error density