## Generalized structured additive regression based on Bayesian P-splines (2006)

### Cached

### Download Links

- [www.stat.uni-muenchen.de]
- [www.statistik.lmu.de]
- [www.stat.uni-muenchen.de]
- [epub.ub.uni-muenchen.de]
- DBLP

### Other Repositories/Bibliography

Venue: | Comput. Statist. Data Anal |

Citations: | 27 - 7 self |

### BibTeX

@ARTICLE{Brezger06generalizedstructured,

author = {Andreas Brezger and Stefan Lang},

title = {Generalized structured additive regression based on Bayesian P-splines},

journal = {Comput. Statist. Data Anal},

year = {2006},

volume = {50},

pages = {967--991}

}

### OpenURL

### Abstract

Generalized additive models (GAM) for modeling nonlinear effects of continuous covariates are now well established tools for the applied statistician. A Bayesian version of GAM’s and extensions to generalized structured additive regression (STAR) are developed. One or two dimensional P-splines are used as the main building block. Inference relies on Markov chain Monte Carlo (MCMC) simulation techniques, and is either based on iteratively weighted least squares (IWLS) proposals or on latent utility representations of (multi)categorical regression models. The approach covers the most common univariate response distributions, e.g. the binomial, Poisson or gamma distribution, as well as multicategorical responses. For the first time, Bayesian semiparametric inference for the widely used multinomial logit model is presented. Two applications on the forest health status of trees and a space-time analysis of health insurance data demonstrate the potential of the approach for realistic modeling of complex problems. Software for the methodology is provided within the public domain package BayesX. Key words: geoadditive models, IWLS proposals, multicategorical response, structured additive predictors, surface smoothing

### Citations

1320 | Generalized Additive Models
- Hastie, Tibshirani
- 1990
(Show Context)
Citation Context ...sian responses. A considerable number of competing approaches is now available for modeling and estimating nonlinear functions of continuous covariates. Prominent examples are smoothing splines (e.g. =-=Hastie and Tibshirani, 1990-=-), local polynomials (e.g. Fan and Gijbels, 1996), regression splines with adaptive knot selection (e.g. Friedman and Silverman, 1989; Friedman, 1991; Stone, Hansen, Kooperberg and Truong, 1997) and P... |

1087 | A Practical Guide to Splines - Boor - 1978 |

863 | The Elements of - Hastie, Tibshirani, et al. - 2009 |

623 | Non-Uniform Random Variate Generation
- Devroye
- 1986
(Show Context)
Citation Context ...icient (regarding computing time). A logit model can be expressed in terms of latent utilities by assuming ɛi ∼ N(0, λi) in (15) with λi = 4ψ 2 i , where ψi follows a Kolmogorov-Smirnov distribution (=-=Devroye, 1986-=-). Hence, ɛi is a scale mixture of normal form with a marginal logistic distribution (Andrews and Mallows, 1974). The main difference to the probit case is that additional parameters λi must be sample... |

516 |
Local polynomial modelling and its applications
- Fan, Gijbels
- 1996
(Show Context)
Citation Context ...roaches is now available for modeling and estimating nonlinear functions of continuous covariates. Prominent examples are smoothing splines (e.g. Hastie and Tibshirani, 1990), local polynomials (e.g. =-=Fan and Gijbels, 1996-=-), regression splines with adaptive knot selection (e.g. Friedman and Silverman, 1989; Friedman, 1991; Stone, Hansen, Kooperberg and Truong, 1997) and P-splines (Eilers and Marx, 1996; Marx and Eilers... |

500 | Computer Solution of Large Sparse Positive Definite Systems - George, Liu - 1981 |

452 | Bayesian analysis of binary and polychotomous response data - Albert, Chib - 1993 |

302 | nd Analysis of Longitudinal Data - Diggle, Heagerty, et al. - 2002 |

277 | Multivariate Statistical Modelling Based on Generalized Linear Models - Fahrmeir, Tutz - 2001 |

233 |
Multivariate adaptive regression splines (with discussion), The Annals of Statistics 19
- Friedman
- 1991
(Show Context)
Citation Context ... examples are smoothing splines (e.g. Hastie and Tibshirani, 1990), local polynomials (e.g. Fan and Gijbels, 1996), regression splines with adaptive knot selection (e.g. Friedman and Silverman, 1989; =-=Friedman, 1991-=-; Stone, Hansen, Kooperberg and Truong, 1997) and P-splines (Eilers and Marx, 1996; Marx and Eilers, 1998). Currently, smoothing based on mixed model representations of GAM’s and extensions is extreme... |

224 | Bayesian image restoration, with two applications in spatial statistics, (with discussion - Besag, York, et al. - 1991 |

189 | R: Semiparametric Regression - Ruppert, Wand, et al. - 2003 |

180 | Statistical models in S
- Hastie
- 1989
(Show Context)
Citation Context ...es well known from the literature as special cases. Examples are generalized additive mixed models (e.g. Lin and Zhang, 1999), geoadditive models (Kammann and Wand, 2003), varying coefficient models (=-=Hastie and Tibshirani, 1993-=-) or geographically weighted regression (Fotheringham, Brunsdon and Charlton, 2002). The approach covers the most common univariate response distributions (Gaussian, binomial, Poisson, gamma) as well ... |

146 |
Scale mixtures of normal distributions
- Andrews, Mallows
- 1974
(Show Context)
Citation Context ...ssuming ɛi ∼ N(0, λi) in (15) with λi = 4ψ 2 i , where ψi follows a Kolmogorov-Smirnov distribution (Devroye, 1986). Hence, ɛi is a scale mixture of normal form with a marginal logistic distribution (=-=Andrews and Mallows, 1974-=-). The main difference to the probit case is that additional parameters λi must be sampled. Holmes and Held (2003) propose to joint update Ui, λi by first drawing from the marginal distribution p(Ui |... |

141 | Polynomial splines and their tensor products in extended linear modeling - Stone, Hansen, et al. - 1997 |

136 | Nonparametric regression using Bayesian variable selection
- Smith, Kohn
- 1996
(Show Context)
Citation Context ...tages, e.g. smoothing parameters can be estimated simultaneously with the regression functions. Bayesian approaches are currently either based on regression splines with adaptive knot selection (e.g. =-=Smith and Kohn, 1996-=-; Denison, Mallick and Smith, 1998; Biller, 2000; Di Matteo, Genovese and Kass, 2001; Biller and Fahrmeir, 2001; Hansen and Kooperberg, 2002), or on smoothness priors (Hastie and Tibshirani (2000) and... |

136 | Bayesian measures of model complexity and fit - Spiegelhalter, Best, et al. - 2002 |

134 | Bayesian computation and stochastic systems (with discussion - Besag, Green, et al. - 1995 |

129 | Efficient simulation from the Multivariate Normal and Student-t distributions subject to linear constraints - Geweke - 1991 |

87 | Geographically Weighted Regression: the analysis of spatially varying relationships - Fotheringham, Brunsdon, et al. - 2002 |

83 |
Flexible parsimonious smoothing and additive modeling (with discussion
- Friedman, Silverman
- 1989
(Show Context)
Citation Context ...ntinuous covariates. Prominent examples are smoothing splines (e.g. Hastie and Tibshirani, 1990), local polynomials (e.g. Fan and Gijbels, 1996), regression splines with adaptive knot selection (e.g. =-=Friedman and Silverman, 1989-=-; Friedman, 1991; Stone, Hansen, Kooperberg and Truong, 1997) and P-splines (Eilers and Marx, 1996; Marx and Eilers, 1998). Currently, smoothing based on mixed model representations of GAM’s and exten... |

76 | On conditional and intrinsic autoregressions - Besag, Kooperberg - 1995 |

75 | Automatic Bayesian Curve Fitting - Denison, Mallick, et al. - 1998 |

74 | Fast sampling of Gaussian Markov random fields - Rue |

68 | Bayesian P-Splines - Lang, Brezger |

66 | Simulation of truncated normal variables - Robert - 1995 |

64 | S (2001) Bayesian Inference for Generalized Additive Mixed Models Based on Markov Random Field Priors - Fahrmeir, Lang |

63 | Efficient sampling from the posterior distribution in generalized linear models - Gamerman - 1997 |

57 | Bayesian Curve-Fitting With Free-Knot Splines - DiMatteo, Genovese, et al. - 2001 |

51 | On block updating in markov random field models for disease mapping - Knorr-Held, Rue |

50 | Computational methods for local regression - Cleveland, Grosse - 1991 |

50 | Modelling and smoothing parameter estimation with multiple quadratic penalties - Wood - 2000 |

46 | Inference in generalized additive mixed models by using smoothing splines
- Lin, Zhang
- 1999
(Show Context)
Citation Context ...C algorithms for posterior inference are developed. STAR models include a number of model classes well known from the literature as special cases. Examples are generalized additive mixed models (e.g. =-=Lin and Zhang, 1999-=-), geoadditive models (Kammann and Wand, 2003), varying coefficient models (Hastie and Tibshirani, 1993) or geographically weighted regression (Fotheringham, Brunsdon and Charlton, 2002). The approach... |

36 |
Direct generalized additive modeling with penalized likelihood. Computational Statistics and Data Analysis
- Marx, Eilers
- 1998
(Show Context)
Citation Context ...d Gijbels, 1996), regression splines with adaptive knot selection (e.g. Friedman and Silverman, 1989; Friedman, 1991; Stone, Hansen, Kooperberg and Truong, 1997) and P-splines (Eilers and Marx, 1996; =-=Marx and Eilers, 1998-=-). Currently, smoothing based on mixed model representations of GAM’s and extensions is extremely popular, see e.g. Lin and Zhang (1999), Currie and Durban (2002), Wand (2003) and the book by Ruppert,... |

35 | Geoadditive models - KAMMANN, P - 2003 |

33 | Conditional prior proposals in dynamic models - Knorr-Held - 1999 |

30 |
Flexible smoothing using B-splines and penalized likelihood
- Eilers, Marx
- 1996
(Show Context)
Citation Context ...olynomials (e.g. Fan and Gijbels, 1996), regression splines with adaptive knot selection (e.g. Friedman and Silverman, 1989; Friedman, 1991; Stone, Hansen, Kooperberg and Truong, 1997) and P-splines (=-=Eilers and Marx, 1996-=-; Marx and Eilers, 1998). Currently, smoothing based on mixed model representations of GAM’s and extensions is extremely popular, see e.g. Lin and Zhang (1999), Currie and Durban (2002), Wand (2003) a... |

30 | Bayesian Semiparametric Regression Analysis of Multicategorical Time–Space Data - Fahrmeir, Lang - 2001 |

30 | Nonparametric regression using linear combinations of basis functions - Kohn, Smith, et al. - 2001 |

29 | Thin plate regression splines - Wood - 2003 |

24 | Smoothing and Mixed Models - Wand - 2003 |

21 | Adaptive Bayesian regression splines in semiparametric generalized linear models - Biller - 2000 |

21 | Analysis of Longitudinal Data - PJ, Liang, et al. - 1996 |

19 | Practical Markov Chain Monte Carlo - Green - 1994 |

19 | Bayesian Analysis for Correlated Ordinal Data Models - Chen, Dey - 2000 |

18 | Bayesian inference for finite mixtures of generalized linear models with random effects - Lenk, DeSarbo - 2000 |

15 | Generalized linear models: A Bayesian perspective - Dey, Ghosh, et al. - 1988 |

13 | Space-varying regression models: specifications and simulations - Gamerman, Moreira, et al. - 2002 |

13 | Bayesian backfitting - Hastie, Tibshirani - 1993 |

12 | Spline adaption in extended linear models - Hansen, Kooperberg - 2002 |