## Polynomial Splines and Their Tensor Products in Extended Linear Modeling (1997)

Venue: | Ann. Statist |

Citations: | 141 - 14 self |

### BibTeX

@ARTICLE{Stone97polynomialsplines,

author = {Charles J. Stone and Mark Hansen and Charles Kooperberg and Young K. Truong},

title = {Polynomial Splines and Their Tensor Products in Extended Linear Modeling},

journal = {Ann. Statist},

year = {1997},

volume = {25},

pages = {1371--1470}

}

### Years of Citing Articles

### OpenURL

### Abstract

ANOVA type models are considered for a regression function or for the logarithm of a probability function, conditional probability function, density function, conditional density function, hazard function, conditional hazard function, or spectral density function. Polynomial splines are used to model the main effects, and their tensor products are used to model any interaction components that are included. In the special context of survival analysis, the baseline hazard function is modeled and nonproportionality is allowed. In general, the theory involves the L 2 rate of convergence for the fitted model and its components. The methodology involves least squares and maximum likelihood estimation, stepwise addition of basis functions using Rao statistics, stepwise deletion using Wald statistics, and model selection using BIC, cross-validation or an independent test set. Publically available software, written in C and interfaced to S/S-PLUS, is used to apply this methodology to...

### Citations

3904 | Classification and Regression Trees - Breiman, Friedman, et al. - 1984 |

2307 | Estimating the dimension of a model - SCHWARZ - 1978 |

2162 | Density Estimation for Statistics and Data Analysis. Monographson Statistics and Applied Probability - Silverman - 1986 |

1489 |
Fundamentals of Speech Recognition
- Rabiner, Juang
- 1993
(Show Context)
Citation Context ...ined. The coefficients of the Fourier transform representation of the log-magnitude of this model are known as its cepstral coefficients. The PLP features (Bourlard and Morgan, 1994; Hermansky, 1990; =-=Rabiner and Juang, 1993-=-) that we used are the log-gain of the model (similar to the variance) and the next eight cepstral coefficients (similar to the autoregressive coefficients). The goal in our analysis is to estimate th... |

1320 | Generalized Additive Models - Hastie, Tibshirani - 1990 |

1272 |
Spline Models for Observational Data
- Wahba
- 1990
(Show Context)
Citation Context ...ed as smoothing, for which there are numerous methods available. Some of the popular methods are kernel and local polynomial regression (Wand and Jones 1995; Fan and Gijbels 1996), smoothing splines (=-=Wahba 1990-=-; Green and Silverman 1994), and polynomial splines. Smith (1982) is probably the first paper to use polynomial splines with adaptively selected knots for regression problems. In her method, knots for... |

1085 | A Pmctical Guide to Splines - Boor - 1978 |

1021 | P.: Nonparametric estimation from incomplete observations - Kaplan, Meier - 1958 |

586 |
Perceptual Linear Predictive (PLP) Analysis of Speech,” The
- Hermansky
- 1990
(Show Context)
Citation Context ... spectrum is obtained. The coefficients of the Fourier transform representation of the log-magnitude of this model are known as its cepstral coefficients. The PLP features (Bourlard and Morgan, 1994; =-=Hermansky, 1990-=-; Rabiner and Juang, 1993) that we used are the log-gain of the model (similar to the variance) and the next eight cepstral coefficients (similar to the autoregressive coefficients). The goal in our a... |

516 |
Local polynomial modelling and its applications
- Fan, Gijbels
- 1996
(Show Context)
Citation Context ...ctor, nonparametric regression can be viewed as smoothing, for which there are numerous methods available. Some of the popular methods are kernel and local polynomial regression (Wand and Jones 1995; =-=Fan and Gijbels 1996-=-), smoothing splines (Wahba 1990; Green and Silverman 1994), and polynomial splines. Smith (1982) is probably the first paper to use polynomial splines with adaptively selected knots for regression pr... |

485 | The statistical analysis of failure times data - Kalbfleisch, Prentice - 1980 |

455 |
Nonparametric Regression and Generalized Linear Models: A roughness penalty approach
- Green, Silverman
- 1994
(Show Context)
Citation Context ...ing, for which there are numerous methods available. Some of the popular methods are kernel and local polynomial regression (Wand and Jones 1995; Fan and Gijbels 1996), smoothing splines (Wahba 1990; =-=Green and Silverman 1994-=-), and polynomial splines. Smith (1982) is probably the first paper to use polynomial splines with adaptively selected knots for regression problems. In her method, knots for cubic 19 splines are posi... |

399 | Time Series: Data Analysis and Theory - Brillinger - 1981 |

399 |
Kernel Smoothing
- M, Jones
- 1995
(Show Context)
Citation Context ...ere is only one predictor, nonparametric regression can be viewed as smoothing, for which there are numerous methods available. Some of the popular methods are kernel and local polynomial regression (=-=Wand and Jones 1995-=-; Fan and Gijbels 1996), smoothing splines (Wahba 1990; Green and Silverman 1994), and polynomial splines. Smith (1982) is probably the first paper to use polynomial splines with adaptively selected k... |

334 | D.,Analysas of survival data - Cox, Oakes - 1984 |

330 | Regression Models and Life Tables (with Discussion - Cox - 1972 |

232 |
Multivariate adaptive regression splines (with discussion
- Friedman
- 1991
(Show Context)
Citation Context ... as "POLYMARS". 5.1 MARS Let (X 1 ; Y 1 ); : : : ; (X n ; Y n ) denote a random sample from the distribution of (X; Y ), where X 2 IR M and Y 2 IR. We wish to estimate f(X) = E(Y jX). The MA=-=RS model (Friedman 1991-=-) can be written as f(X) = f(Xjfi) = J X j=1 fi j B j (X): (5.1) 20 For a given set of basis functions, the unknown parameters in MARS are estimated using least squares. The selection of the basis fun... |

218 | A method to identify protein sequences that fold into a known three dimensional structure - Bowie, Luthy, et al. - 1991 |

158 | Additive Regression and Other Nonparametric Models - Stone - 1985 |

129 | Continuous speech recognition
- N, Bourlard
- 1995
(Show Context)
Citation Context ...le model of the transformed spectrum is obtained. The coefficients of the Fourier transform representation of the log-magnitude of this model are known as its cepstral coefficients. The PLP features (=-=Bourlard and Morgan, 1994-=-; Hermansky, 1990; Rabiner and Juang, 1993) that we used are the log-gain of the model (similar to the variance) and the next eight cepstral coefficients (similar to the autoregressive coefficients). ... |

117 | Counting processes and survival analysis - Fleming, Harrington - 1991 |

114 | Multivariate Splines - Chui - 1988 |

112 | Flexible discriminant analysis by optimal scoring - Hastie, Tibshirani, et al. - 1994 |

94 | Covariance analysis of censored survival data - Breslow - 1974 |

83 | Flexible parsimonious smoothing and additive modeling (with discussion - Friedman, Silverman - 1989 |

82 | Selection of representative protein datasets - Hobohm, Scharf, et al. - 1992 |

80 | Hazard regression - Kooperberg, Stone, et al. - 1995 |

77 | Triangular Bernstein-Bézier patches - Farin - 1986 |

59 | The dimensionality reduction principle for generalized additive models - Stone - 1986 |

57 | Survival analysis - Miller, Gong, et al. - 1981 |

54 | Statistical Models Based on Counting Processes - Gill, Keiding - 1993 |

52 | Regression Diagnostics: Identifiing Influential Data and Sources of Collinearity - Belsey, Kuh, et al. - 1980 |

47 | Logspline density estimation for censored data - Kooperberg, Stone - 1992 |

44 | Smoothing spline ANOVA with component-wise Bayesian “confidence intervals - Gu, C, et al. - 1993 |

43 | A Study of Logspline Density Estimation - Kooperberg, Stone - 1991 |

41 | Flexible methods for analyzing survival data using splines, with application to breast cancer prognosis - Gray - 1992 |

39 | Statistical Computing - Kennedy, Gentle - 1980 |

35 | Nonparametric statistical data modeling - Parzen - 1979 |

32 | B-form basics - Boor - 1987 |

28 | Speaker-Independent Continuous Speech Dictation
- Gauvain, Lamel, et al.
- 1993
(Show Context)
Citation Context ...e of the text of the utterance. The phoneme transcription, which we obtained from the International Computer Science Institute (ICSI) in Berkeley, California, is based on the LIMSI phonetic alphabet (=-=Gauvain et al. 1994-=-). The utterances were also processed to produce perceptual linear predictive (PLP) features. Every 12.5 ms the audible spectrum, based on a concentric 25 ms piece of sound, is determined. Since we co... |

24 | Additive Splines in Statistics - Stone, Koo - 1986 |

20 | Large-sample inference for log-spline models - Stone - 1990 |

15 | A telephone speech database of spelled and spoken names - Cole, Roginski, et al. |

14 | Logspline estimation of a possibly mixed spectral distribution - Kooperberg, Stone, et al. - 1995 |

13 | Triogram models - Hansen, Kooperberg, et al. - 1998 |

13 | Curve fitting and modeling with splines using statistical variable selection techniques - Smith - 1982 |

11 | Regression splines in the Cox model with application to covariate effects in liver disease - Sleeper, Harrington - 1990 |

10 | Corpus development activities at the Center for Spoken Language Understanding - Cole, Noel, et al. - 1994 |

10 | The L2 rate of convergence for hazard regression - Kooperberg, Stone, et al. - 1995 |

9 | Discussion of the paper by - Breslow - 1972 |