## Generalized Partially Linear Single-Index Models (1998)

Venue: | Journal of the American Statistical Association |

Citations: | 80 - 27 self |

### BibTeX

@ARTICLE{Carroll98generalizedpartially,

author = {R. J. Carroll and Jianqing Fan and Irène Gijbels and M.P. Wand},

title = {Generalized Partially Linear Single-Index Models},

journal = {Journal of the American Statistical Association},

year = {1998},

volume = {92},

pages = {477--489}

}

### Years of Citing Articles

### OpenURL

### Abstract

The typical generalized linear model for a regression of a response Y on predictors (X; Z) has conditional mean function based upon a linear combination of (X; Z). We generalize these models to have a nonparametric component, replacing the linear combination T 0 X + T 0 Z by 0 ( T 0 X) + T 0 Z, where 0 ( ) is an unknown function. We call these generalized partially linear single-index models (GPLSIM). The models include the "single-index" models, which have 0 = 0. Using local linear methods, estimates of the unknown parameters ( 0 ; 0 ) and the unknown function 0 ( ) are proposed, and their asymptotic distributions obtained. Examples illustrate the models and the proposed estimation methodology.

### Citations

1614 | A simplex method for function minimization - Nelder, Mead - 1964 |

1575 | Generalized additive models
- Hastie, Tibshirani
- 1990
(Show Context)
Citation Context ..., one must undersmooth the nonparametric part 0 (). The need to undersmooth to obtain usual rates of convergence is standard in the kernel 14 literature and has analogues in the spline literature (Ha=-=stie and Tibshirani, 19-=-90, pp. 154-155). This undersmoothing is required for the estimator dened by (9). However, for the estimator dened by (10), in the linear regression single-index model with no Z, ordinary bandwidth ra... |

724 |
Robust Locally Weighted Regression and Smoothing Scatterplots
- Cleveland
- 1979
(Show Context)
Citation Context ...rmine smoothing neighborhoods. For example, one might use a locally determined bandwidth or even a nearest neighbor scheme to determine the smoothing neighborhoods, similar to what is done in lowess (=-=Cleveland, 1979-=-). In principle, there is no need to restrict to kernel estimation of any type, with splines being a natural alternative. It would be interesting to investigate all of these possibilities. ?From a the... |

294 | Design Adaptive Nonparametric Regression - Fan - 1992 |

234 |
Generalized Linear Models, Second Edition
- McCullagh, Nelder
- 1989
(Show Context)
Citation Context ...al variance is modeled as var(Y jX = x; Z = z) = 2 V f(x; z)g for some known positive function V , then the corresponding quasilikelihood function Q(w; y) satises @ @w Q(w; y) = (y w)=V (w): (7) (McC=-=ullagh and Nelder 1989-=-, Chapter 9). The quasiscore (7) possesses properties similar to those of the usual likelihood score function. In Section 2, we propose estimation procedures, and their performance is illustrated in s... |

218 | Sliced Inverse Regression for Dimension Reduction” (with discussions - Li - 1991 |

143 | An effective bandwidth selector for local least squares regression (Corr: 96V91 p1380 - Ruppert, Sheather, et al. - 1995 |

131 |
Quasi-likelihood functions, generalized linear models, and the Gauss-Newton method
- Wedderburn
- 1974
(Show Context)
Citation Context ...e is modeled as var(Y jX = x; Z = z) = oe 2 V f��(x; z)g for some known positive function V , then the corresponding quasilikelihood function Q(w; y) satisfies @ @w Q(w; y) = (y \Gamma w)=V (w): (=-=7) (Wedderburn 1974-=-, McCullagh & Nelder 1989, Chapter 9). The quasiscore (7) possesses properties similar to those of the usual likelihood score function. 1.3 Outline In Section 2, we discuss the algorithms used to esti... |

93 | Kernel smoothing in partial linear models - Speckman - 1988 |

86 | Weak and strong uniform consistency of kernel regression estimates - Mack, Silverman - 1982 |

67 | 1994a�. Fast implementations of nonparametric curve estimators
- Fan, Marron
(Show Context)
Citation Context ... covariance matrix is e H 2 e A 1 e H T 2 . 18 8 IMPLEMENTATION To cut down on the computational labor at the curve estimation stages fast binned approximations are used (e.g. Hardle and Scott 1992, F=-=an and Marron 1994-=-). Binning methods can also be used for fast computation of the standard error estimates. Details of such calculations are given in Turlach and Wand (1995). An S-PLUS/Fortran module forstting GPLSIM i... |

65 | Quasi-likelihood estimation in semiparametric models - Severini, Staniswalis - 1994 |

65 |
Profile likelihood and conditionally parametric models
- Severini, Wong
- 1992
(Show Context)
Citation Context ...e as shown by Hardle, Hall and Ichimura (1993), who suggest that one maximize (10) simultaneously in the bandwidth and the parameters. Severini and Staniswalis (1994) and Hunsberger (1994) (see also S=-=everini and Wong, 1992) sho-=-w the same thing for the partially linear model. Since ordinary bandwidths \work" for single-index models and also for partially linear models, it is reasonable to suppose that they work for the ... |

62 | Local polynomial kernel regression for generalized linear models and quasi-likelihood functions - Fan, Heckman, et al. - 1995 |

61 | Asymptotics for Least Absolute Deviation Regression Estimators", Econometric Theory - Pollard - 1991 |

45 | Efficient and Adaptive Inference for Semiparametric Models - Bickel, Klaasen, et al. - 1993 |

40 | Optimal smoothing in single-index models - Härdle, Hall, et al. - 1993 |

38 | Convergence rates for parametric components in a partly linear - Chen - 1988 |

37 | Spline smoothing in partly linear models - Heckman - 1986 |

35 | Generalized Additive Models: Some Applications - Hastie, Tibshirani - 1987 |

31 | Robust reconstruction of functions by local-approximation method - Tsybakov - 1986 |

26 | Smoothing by weighted averaging of rounded points - Härdle, Scott - 1992 |

20 | Partial spline models for semiparametric estimation of functions of several variables - Wahba, G - 1984 |

18 | Semiparametric additive regression - Cuzick - 1992 |

18 | A statistical method for assessing a threshold in epidemiological studies. Stat Med - Ulm - 1991 |

13 | Generalized linear models - Wedderburn - 1972 |

12 |
Overall and Coronary Heart Disease Mortality Rates in Relation to Major Risk Factors in 325,348 men Screened for MRFIT
- Kannel, Newton, et al.
- 1986
(Show Context)
Citation Context ...erche Concertee' (No. 93-98-164). We thank the referee and associate editor for many helpful comments which signicantly improved the paper. 1 INTRODUCTION 1.1 Motivation The Framingham Heart Study (Ka=-=nnel, et al., 1986-=-) consists of a series of exams taken two years apart. For purposes of illustration, we use Exam #3 as the baseline. There are 1,615 men aged 31-65 in this data set, with the outcome indicating the oc... |

11 | Semiparametric regression in likelihood-based models - Hunsberger - 1994 |

10 | Segmented regression with errors in predictors: semiparametric and parametric methods - Küchenhoff, Carroll - 1997 |

5 | Fast computation of auxiliary quantities in local polynomial regression - Turlach, Wand - 1995 |

3 | Semiparametric estimation of generalized linear models and related models," Manuscript - Bonneu, Delecroix, et al. - 1995 |

3 | Estimating the Missing Link Function - Weisberg, Welsh - 1994 |

1 | Penalized estimation in partial linear models - Mammen, Geer - 1995 |

1 | Regression Using Fractional Polynomials of Continuous Covariates: Parsimonious Parametric Modelling - P, Altman - 1994 |

1 | p and f(x; ) is a continuous function in 2 C and x 2 D. Assume that b (x) 2 C is continuous in x 2 D, and is the unique maximizer of f(x; ). Let b n (x) 2 C be a maximizer of f n (x; ). If sup 2C;x2D jf n (x; ) f(x; )j ! 0; then sup x2D j b n (x) b (x) - IR - 1982 |

1 | Estimating the Missing Link Function. Annals of Statistics, to appear. 9 APPENDIX 9.1 Proof of Theorem 1 Here we prove the results for the one-step estimators. The methods of technique are similar for the fully iterated estimators. For simplicity of notat - Welsh - 1994 |