Semiparametric regression

In statistics, semiparametric regression includes regression models that combine parametric and nonparametric models. They are often used in situations where the fully nonparametric model may not perform well or when the researcher wants to use a parametric model but the functional form with respect to a subset of the regressors or the density of the errors is not known. Semiparametric regression models are a particular type of semiparametric modelling and, since semiparametric models contain a parametric component, they rely on parametric assumptions and may be misspecified and inconsistent, just like a fully parametric model.

Methods

edit

Many different semiparametric regression methods have been proposed and developed. The most popular methods are the partially linear, index and varying coefficient models.

Partially linear models

edit

A partially linear model is given by

 

where   is the dependent variable,   is a   vector of explanatory variables,   is a   vector of unknown parameters and  . The parametric part of the partially linear model is given by the parameter vector   while the nonparametric part is the unknown function  . The data is assumed to be i.i.d. with   and the model allows for a conditionally heteroskedastic error process   of unknown form. This type of model was proposed by Robinson (1988) and extended to handle categorical covariates by Racine and Li (2007).

This method is implemented by obtaining a   consistent estimator of   and then deriving an estimator of   from the nonparametric regression of   on   using an appropriate nonparametric regression method.[1]

Index models

edit

A single index model takes the form

 

where  ,   and   are defined as earlier and the error term   satisfies  . The single index model takes its name from the parametric part of the model   which is a scalar single index. The nonparametric part is the unknown function  .

Ichimura's method

edit

The single index model method developed by Ichimura (1993) is as follows. Consider the situation in which   is continuous. Given a known form for the function  ,   could be estimated using the nonlinear least squares method to minimize the function

 

Since the functional form of   is not known, we need to estimate it. For a given value for   an estimate of the function

 

using kernel method. Ichimura (1993) proposes estimating   with

 

the leave-one-out nonparametric kernel estimator of  .

Klein and Spady's estimator

edit

If the dependent variable   is binary and   and   are assumed to be independent, Klein and Spady (1993) propose a technique for estimating   using maximum likelihood methods. The log-likelihood function is given by

 

where   is the leave-one-out estimator.

Smooth coefficient/varying coefficient models

edit

Hastie and Tibshirani (1993) propose a smooth coefficient model given by

 

where   is a   vector and   is a vector of unspecified smooth functions of  .

  may be expressed as

 

See also

edit

Notes

edit
  1. ^ See Li and Racine (2007) for an in-depth look at nonparametric regression methods.

References

edit
  • Robinson, P.M. (1988). "Root-n Consistent Semiparametric Regression". Econometrica. 56 (4). The Econometric Society: 931–954. doi:10.2307/1912705. JSTOR 1912705.
  • Li, Qi; Racine, Jeffrey S. (2007). Nonparametric Econometrics: Theory and Practice. Princeton University Press. ISBN 978-0-691-12161-1.
  • Racine, J.S.; Qui, L. (2007). "A Partially Linear Kernel Estimator for Categorical Data". Unpublished Manuscript, Mcmaster University.
  • Ichimura, H. (1993). "Semiparametric Least Squares (SLS) and Weighted SLS Estimation of Single Index Models". Journal of Econometrics. 58 (1–2): 71–120. doi:10.1016/0304-4076(93)90114-K.
  • Klein, R. W.; R. H. Spady (1993). "An Efficient Semiparametric Estimator for Binary Response Models". Econometrica. 61 (2). The Econometric Society: 387–421. CiteSeerX 10.1.1.318.4925. doi:10.2307/2951556. JSTOR 2951556.
  • Hastie, T.; R. Tibshirani (1993). "Varying-Coefficient Models". Journal of the Royal Statistical Society, Series B. 55: 757–796.