Abstract:Representation of generalized additive models (GAM's) using penalized regression splines allows GAM's to be employed in a straightforward manner using penalized regression methods. Not only is inference facilitated by this approach, but it is also possible to integrate model selection in the form of smoothing parameter selection into model fitting in a computationally efficient manner using well founded criteria such as generalized cross-validation. The current fitting and smoothing parameter selection methods for such models are usually effective, but do not provide the level of numerical stability to which users of linear regression packages, for example, are accustomed. In particular the existing methods cannot deal adequately with numerical rank deficiency of the GAM fitting problem, and it is not straightforward to produce methods that can do so, given that the degree of rank deficiency can be smoothing parameter dependent. In addition, models with the potential flexibility of GAM's can also present practical fitting difficulties as a result of indeterminacy in the model likelihood: Data with many zeros fitted by a model with a log link are a good example. In this article it is proposed that GAM's with a ridge penalty provide a practical solution in such circumstances, and a multiple smoothing parameter selection method suitable for use in the presence of such a penalty is developed. The method is based on the pivoted QR decomposition and the singular value decomposition, so that with or without a ridge penalty it has good error propagation properties and is capable of detecting and coping elegantly with numerical rank deficiency. The method also allows mixtures of user specified and estimated smoothing parameters and the setting of lower bounds on smoothing parameters. In terms of computational efficiency, the method compares well with existing methods. A simulation study compares the method to existing methods, including treating GAM's as mixed models.

Component Selection in the Additive Regression Model

Variable Selection for Additive Model Via Cumulative Ratios of Empirical Strengths Total

Variable Selection for Generalized Varying Coefficient Partially Linear Models with Diverging Number of Parameters

Two-step estimation of high dimensional additive models

Statistical Inference for Generalized Additive Models: Simultaneous Confidence Corridors and Variable Selection

Additive model building for spatial regression

Model selection and estimation in regression with grouped variables

Robust variable selection for partially linear additive models

Estimation and model selection in generalized additive partial linear models for correlated data with diverging number of covariates

PENALIZED VARIABLE SELECTION PROCEDURE FOR COX MODELS WITH SEMIPARAMETRIC RELATIVE RISK

SPLINE-BACKFITTED KERNEL SMOOTHING OF ADDITIVE COEFFICIENT MODEL

Globally consistent model selection in semi-parametric additive coefficient models

ADDITIVE COEFFICIENT MODELING VIA POLYNOMIAL SPLINE

Variable selection for semiparametric varying-coefficient spatial autoregressive models with a diverging number of parameters

Identification of partially linear structure in additive models with an application to gene expression prediction from sequences

A one-covariate-at-a-time multiple testing approach to variable selection in additive models

Identification of Non-Linear Additive Autoregressive Models

Stable and Efficient Multiple Smoothing Parameter Estimation for Generalized Additive Models

Model Selection for Exponential Power Mixture Regression Models

A robust spline approach in partially linear additive models