Abstract
In the framework of quantile regression, local linear smoothing techniques have been studied by several authors, particularly by Yu and Jones (J Am Stat Assoc 93:228–237, 1998). The problem of bandwidth selection was addressed in the literature by the usual approaches, such as crossvalidation or plugin methods. Most of the plugin methods rely on restrictive assumptions on the quantile regression model in relation to the mean regression, or on parametric assumptions. Here we present a plugin bandwidth selector for nonparametric quantile regression that is defined from a completely nonparametric approach. To this end, the curvature of the quantile regression function and the integrated squared sparsity (inverse of the conditional density) are both nonparametrically estimated. The new bandwidth selector is shown to work well in different simulated scenarios, particularly when the conditions commonly assumed in the literature are not satisfied. A real data application is also given.
This is a preview of subscription content, log in to check access.
References
Abberger K (1998) Crossvalidation in nonparametric quantile regression. Allg Statistisches Archiv 82:149–161
Abberger K (2002) Variable data driven bandwidth choice in nonparametric quantile regression. Technical Report
Bloch DA, Gastwirth JL (1968) On a simple estimate of the reciprocal of the density function. Ann Math Stat 39:1083–1085
Bofinger E (1975) Estimation of a density function using order statistics. Aust J Stat 17:1–7
CondeAmboage M (2017) Statistical inference in quantile regression models, Ph.D. Thesis. Universidade de Santiago de Compostela, Spain
El Ghouch A, Genton MG (2012) Local polynomial quantile regression with parametric features. J Am Stat Assoc 104:1416–1429
Fan J, Hu TC, Truong YK (1994) Robust nonparametric function estimation. Scand J Stat 21:433–446
Hall P, Sheather SJ (1988) On the distribution of a studentized quantile. J R Stat Soc Ser B (Methodol) 50:381–391
Jones MC (1991) The roles of ISE and MISE in density estimation. Stat Probab Lett 12:51–56
Jones MC, Yu K (2007) Improved double kernel local linear quantile regression. Stat Model 7:377–389
Koenker R (2005) Quantile regression. Cambridge University Press, Cambridge
Koenker R, Bassett G (1978) Regression quantiles. Econometrica 46:33–50
Mallows CL (1973) Some comments on \(C_p\). Technometrics 15:661–675
Nelder JA, Mead R (1965) A simplex algorithm for function minimization. Comput J 7:308–313
Opsomer JD, Ruppert D (1998) A fully automated bandwidth selection method for fitting additive models. J Am Stat Assoc 93:605–618
Ruppert D, Sheather SJ, Wand MP (1995) An effective bandwidth selector for local least squares regression. J Am Stat Assoc 90:1257–1270
SánchezSellero C, GonzálezManteiga W, Cao R (1999) Bandwidth selection in density estimation with truncated and censored data. Ann Inst Stat Math 51:51–70
Sheather SJ, Maritz JS (1983) An estimate of the asymptotic standard error of the sample median. Aust J Stat 25:109–122
Siddiqui MM (1960) Distribution of quantiles in samples from a bivariate population. J Res Natl Bureau Stand 64:145–150
Tukey JW (1965) Which part of the sample contains the information. Proc Natl Acad Sci 53:127–134
Venables WN, Ripley BD (1999) Modern applied statistics with SPLUS, 3rd edn. Springer, New York
Yu K, Jones MC (1998) Local linear quantile regression. J Am Stat Assoc 93:228–237
Yu K, Lu Z (2004) Local linear additive quantile regression. Scand J Stat 31:333–346
Acknowledgements
The authors gratefully acknowledge the support of Projects MTM2013–41383–P (Spanish Ministry of Economy, Industry and Competitiveness) and MTM2016–76969–P (Spanish State Research Agency, AEI), both cofunded by the European Regional Development Fund (ERDF). Support from the IAP network StUDyS, from Belgian Science Policy, is also acknowledged. Work of M. CondeAmboage has been supported by FPU grant AP20125047 from the Spanish Ministry of Education. We are grateful to two anonymous referees for their constructive comments, which helped to improve the paper.
Author information
Affiliations
Corresponding author
Appendix: Mean squared error of curvature and sparsity estimators
Appendix: Mean squared error of curvature and sparsity estimators
Here expressions (4) and (5) are derived. They give approximations to the mean squared error of curvature and sparsity estimators, respectively. A complete development of these expressions can be seen in Chapter 3 of CondeAmboage (2017).
Derivation of (4)
In order to derive the asymptotic mean integrated squared error of the curvature estimator, the following assumptions will be needed:
 C1:

The density function of the explanatory variable X, denoted by g, is differentiable, and its first derivative is a bounded function.
 C2:

The kernel function K is symmetric and nonnegative, has a bounded support and verifies that \(\int K(u) \; \hbox {d}u =1\), \(\mu _6(K)=\int u^6 K(u) \; \hbox {d}u < \infty \) and \(\int K^2(u) \; \hbox {d}u < \infty \). Moreover, it is assumed that the bandwidth parameter \(h_\mathrm{c}\) verifies that \(h_\mathrm{c} \rightarrow 0\) and \(nh_\mathrm{c}^{5} \rightarrow \infty \) when \(n \rightarrow \infty \).
 C3:

The conditional distribution function \(F(yX=x)\) of the response variable is three times derivable in x for each y and its first derivative verifies that \(F^{(1)}(q_{\tau }(x)X=x)=f(q_{\tau }(x)X=x)\ne 0\). Moreover, there exist positive constants \(c_1\) and \(c_2\) and a positive function \(\text{ Bound }(yX=x)\) such that
$$\begin{aligned} \sup _{x_nx<c_1} f(yX=x_n) \le \text{ Bound }(yx) \end{aligned}$$and
$$\begin{aligned}&\int \psi _{\tau }(yq_{\tau }(x))^{2+\delta } \; \text{ Bound }(yX=x) \; \hbox {d}y<\infty \\&\int (\rho _{\tau }(yt)\rho _{\tau }(y)\psi _{\tau }(y)t)^{2} \; \hbox { Bound}(yX=x) \; \hbox {d}y=o(t^2), \quad \text{ as }\ t \rightarrow 0 \end{aligned}$$where \(\psi _{\tau }(r)=\tau {\mathbb {I}}(r>0)+(\tau 1){\mathbb {I}}(r<0)\).
 C4:

The function \(q_{\tau _1}(x)\) has a continuous fourth derivative with respect to x for any \(\tau _1\) in a neighbourhood of \(\tau \). These derivatives will be denoted by \(q_{\tau }^{(i)}\) with \(i \in \{1,2,3,4\}\). Moreover, all these derivatives are bounded functions in a neighbourhood of \(\tau \).
Applying the arguments of the proof of Theorem 3 in Fan et al. (1994) to a local polynomial of order 3, the estimator of the second derivative can be approximated by
where
where \(Y_i^{(3)}=Y_iq_\tau (x)q_\tau ^{(1)}(x)(X_ix)(1/2)q_\tau ^{(2)}(x)(X_ix)^2(1/6)q_\tau ^{(3)}(x)(X_ix)^3\) and \(\psi _\tau (z)=\tau {\mathbb {I}}(z<0)\). Note that assumptions C1C4 were used here.
Now, expectation and variance of \({\widetilde{q}}_{\tau ,h_\mathrm{c}}^{(2)}(x)\) can be obtained by some algebraic calculations:
where \(\delta _1\) and \(\delta _2\) were defined in expression (4). Recall that the curvature estimator is given by
Then, combining expectation and variance of \({\widetilde{q}}_{\tau ,h_\mathrm{c}}^{(2)}(X_i)\) conditionally to \(X_i\), and taking expectation with respect to \(X_i\), we obtain
Additional calculations, which can be found in CondeAmboage (2017), show that the dominant terms in the variance of \({\widehat{\vartheta }}_{h_\mathrm{c}}\) are of orders \(n^{1}\) and \(n^{2} \, h_\mathrm{c}^{9}\). The term of order \(n^{1}\) does not depend on \(h_\mathrm{c}\), while the term of order \(n^{2}\,h_\mathrm{c}^{9}\) is negligible with respect to the asymptotic squared bias. Because of this, the asymptotically optimal bandwidth can be obtained by minimizing the asymptotic squared bias. This fact, together with last expression for \({\mathbb {E}}\left( {\widehat{\vartheta }}_{h_\mathrm{c}}\right) \), leads to expression (4).
Derivation of (5)
The following conditions will be assumed in order to derive the asymptotic mean integrated squared error of the sparsity estimator:
 S1:

The conditional density function \(f(yX=x)\) of the response variable is twice derivable in x for each y and \(f^{(i)}(q_{\tau }(x)X=x)\ne 0\) with \(i=0,1,2\). Moreover, there exist positive constants \(c_1\) and \(c_2\) and a positive function \(\text{ Bound }(yX=x)\) such that
$$\begin{aligned} \sup _{x_nx<c_1} f(yX=x_n) \le \text{ Bound }(yX=x) \end{aligned}$$and
$$\begin{aligned}&\int \psi _{\tau }(yq_{\tau }(x))^{2+\delta } \; \text{ Bound }(yX=x) \; \hbox {d}y<\infty \\&\\&\int (\rho _{\tau }(yt)\rho _{\tau }(y)\psi _{\tau }(y)t)^{2} \; \hbox { Bound}(yX=x) \; \hbox {d}y=o(t^2), \quad \text{ as }\ t \rightarrow 0 \end{aligned}$$where \(\psi _{\tau }(r)=\tau {\mathbb {I}}(r>0)+(\tau 1){\mathbb {I}}(r<0)\).
 S2:

The function \(q_{\tau _1}\) has a continuous second derivative for any \(\tau _1\) in a neighbourhood of \(\tau \) as a function of x. These derivatives will be denoted by \(q_{\tau }^{(i)}\). Moreover, all these functions are bounded functions in a neighbourhood of \(\tau \).
 S3:

The density function of the explanatory variable X, denoted by g, is differentiable, and this first derivative is a bounded function.
 S4:

The kernel K is symmetric and nonnegative, has a bounded support and verifies that \(\int K(u) \; \hbox {d}u< \infty \), \(\int K(u)^2 \; \hbox {d}u < \infty \) and \(\mu _2(K)<\infty \). Moreover, the bandwidth parameters verify that \(d_\mathrm{s} \rightarrow 0\), \(h_\mathrm{s} \rightarrow 0\) and \(nd_\mathrm{s}h_s \rightarrow \infty \) when \(n \rightarrow \infty \).
 S5:

The function \(q_{\tau _1}\) has a continuous and bounded forth derivative with respect to \(\tau _1\) for any \(\tau _1\) in a neighbourhood of \(\tau \). Moreover, \(q_{\tau _1}^{(2)}\) has a continuous and bounded second derivative with respect to \(\tau _1\) for any \(\tau _1\) in a neighbourhood of \(\tau \).
Recall the definition of the proposed sparsity estimator
where \({\widehat{q}}_{\tau +d_\mathrm{s},h_\mathrm{s}}\) and \({\widehat{q}}_{\tau d_\mathrm{s},h_\mathrm{s}}\) are local linear quantile regression estimates at the quantile orders \((\tau +d_\mathrm{s})\) and \((\tau d_\mathrm{s})\), respectively, and \(h_\mathrm{s}\) denotes their bandwidth. Applying Fan et al. (1994)’s results, we have
where
and \(Y_i^{(1)}=Y_iq_{\tau +d_\mathrm{s}}(x)q_{\tau +d_\mathrm{s}}^{(1)}(x)(X_ix)\). Analogously for \({\widehat{q}}_{\tau d_\mathrm{s},h_\mathrm{s}}(x)\).
Substituting these expressions in the definition of \({\widehat{s}}_{\tau ,d_\mathrm{s},h_\mathrm{s}}(x)\), we have
with
Note that A(x) is not random and can be approximated by a Taylor expansion as
if S2 follows. Moreover, based on arguments developed in Lemma 2 of Fan et al. (1994), the expectation and variance of B(x) can be approximated by
if assumptions S1S5 follow.
From these results, the asymptotic bias of the estimated squared sparsity is given by
where a(x), b(x) and c(x) are given in (6).
In view of expression (11), the asymptotic variance of sparsity estimator can be decomposed as follows
Each of the previous terms can be expressed as covariances of Uexpressions like that given in (10) evaluated at different points x and quantiles \(\tau +d_\mathrm{s}\) and \(\tau d_\mathrm{s}\). These covariances can be computed (under assumptions S1, S2, S4 and S5) using similar arguments to those employed by Fan et al. (1994), adapting their \(\varphi \) function (given in equation (2.1) on p. 435) to each covariance. Then, the asymptotic variance of sparsity estimator can be approximated as follows
where d(x) and e(x) are given in (6). Then, in view of the computed asymptotic bias and variance, expression (5) can be derived.
Rights and permissions
About this article
Cite this article
CondeAmboage, M., SánchezSellero, C. A plugin bandwidth selector for nonparametric quantile regression. TEST 28, 423–450 (2019). https://doi.org/10.1007/s1174901805826
Received:
Accepted:
Published:
Issue Date:
Keywords
 Quantile regression
 Bandwidth
 Nonparametric regression
Mathematics Subject Classification
 62G08