Local and global asymptotic inference in smoothing spline models

Abstract

This article studies local and global inference for smoothing spline estimation in a unified asymptotic framework. We first introduce a new technical tool called functional Bahadur representation, which significantly generalizes the traditional Bahadur representation in parametric models, that is, Bahadur [Ann. Inst. Statist. Math. 37 (1966) 577-580]. Equipped with this tool, we develop four interconnected procedures for inference: (i) pointwise confidence interval; (ii) local likelihood ratio testing; (iii) simultaneous confidence band; (iv) global likelihood ratio testing. In particular, our confidence intervals are proved to be asymptotically valid at any point in the support, and they are shorter on average than the Bayesian confidence intervals proposed by Wahba [J. R. Stat. Soc. Ser. B Stat. Methodol. 45 (1983) 133-150] and Nychka [J. Amer. Statist. Assoc. 83 (1988) 1134-1143]. We also discuss a version of the Wilks phenomenon arising from local/global likelihood ratio testing. It is also worth noting that our simultaneous confidence bands are the first ones applicable to general quasi-likelihood models. Furthermore, issues relating to optimality and efficiency are carefully addressed. As a by-product, we discover a surprising relationship between periodic and nonperiodic smoothing splines in terms of inference.

Published in at http://dx.doi.org/10.1214/13-AOS1164 the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)

References in corpus (1)

• Likelihood based inference for monotone response models

Cited by in corpus (20)

• Nonparametric inference in generalized functional linear models
• Joint asymptotics for semi-nonparametric regression models with partially linear structure
• Solar Flare Index Prediction Using SDO/HMI Vector Magnetic Data Products with Statistical and Machine Learning Methods
• Heteroskedastic and Imbalanced Deep Learning with Adaptive Regularization
• Computational Limits of A Distributed Algorithm For Smoothing Spline
• On Function Approximation in Reinforcement Learning: Optimism in the Face of Large State Spaces
• Nonparametric Heterogeneity Testing For Massive Data
• A Partially Linear Framework for Massive Heterogeneous Data
• Nonparametric Testing under Random Projection
• Non-asymptotic theory for nonparametric testing
• How Many Machines Can We Use in Parallel Computing for Kernel Ridge Regression?
• Deterministic Inequalities for Smooth M-estimators
• Nonparametric Bayesian Aggregation for Massive Data
• Gaussian Approximation of General Nonparametric Posterior Distributions
• Statistical Inference on Panel Data Models: A Kernel Ridge Regression Method
• Kernel-based Partial Permutation Test for Detecting Heterogeneous Functional Relationship
• Sparse and Efficient Estimation for Partial Spline Models with Increasing Dimension
• Statistical Inference on Partially Linear Panel Model under Unobserved Linearity
• Nonparametric Inference under B-bits Quantization
• Minimax Nonparametric Two-sample Test under Smoothing