Abstract
The paper studies a linear errors-in-variables model with first order autoregressive processes. The Huber-Dutter (HD) estimators of unknown parameters are given, and the asymptotic normality of the HD estimators is investigated. Finally, a simple example is given to illustrate our estimation method.
MSC:60F05, 60G10, 62F35, 62M10, 60G42.
Similar content being viewed by others
1 Introduction
Consider the following linear errors-in-variables (EV) model:
where the superscript T denotes the transpose throughout the paper, are scalar response variables, and are observable and unobservable random variables, respectively, is a vector of d unknown parameters, are independent and identically distributed (i.i.d.) measurement errors with and , and are independent, and are independent, and are the first order autoregressive (AR(1)) processes
where are i.i.d. random errors with zero mean and finite variance , and is a one-dimensional unknown parameter. A common assumption is that the ratio of the error variances is known. This is assumed throughout this paper and all variables are assumed scaled so that .
The linear errors-in-variables model (1.1) with AR(1) processes (1.2) includes three important special models: (1) an ordinary linear regression model with AR(1) processes (when , see e.g., Hu [1], Maller [2], Pere [3], and Fuller [4]); (2) an ordinary linear errors-in-variables model (when , see e.g., Miao and Liu [5], Miao et al. [6, 7], Liu and Chen [8], Cui [9], Cui and Chen [10], Cheng and Van Ness [11]); (3) autoregressive processes (when , see e.g., Hamilton [12], Brockwell and Davis [13], and Fuller [4]). The independence assumption for the errors is not always valid in applications, especially for sequentially collected economic data, which often exhibit evident dependence in the errors. Recently, linear errors-in-variables models with serially correlated errors have attracted increasing attention from statisticians; see, for example, Baran [14], Fan et al. [15], Miao et al. [16], among others.
It is well known that in the EV model, the ordinary least-squares (OLS) estimators are biased and inconsistent and that orthogonal regression is better in that case Fuller [17]. However, both methods are very sensitive to outliers in the data and some robust alternatives have been proposed. Brown [18] and Ketellapper and Ronner [19] applied robust ordinary regression techniques in the EV model. Zamar [20] proposed robust orthogonal regression M-estimators and showed that it outperformed the robust ordinary regression. Cheng and Van Ness [21] generalized the proposal of Zamar by defining robust orthogonal Generalized M-estimators which had bounded influence function in the simple case. He and Liang [22] proposed a regression quantile approach in the EV model which allowed for heavier tailed errors distribution than the gaussian distribution. Fekri and Ruiz-Gazen [23] proposed robust weighted orthogonal regression.
Over the last 40 years, several estimators in linear regression models that posses robustness have been proposed, such as Wu [24], Silvapullé [25], Hampel et al. [26], Huber and Ronchetti [27], Li [28], Babu [29], Cheng and Van Ness [21], Salibian-Barrera and Zamar [30], Wu and Wang [31], Zhou and Wu [32], and so on. It is well known that HD estimate approach is one of important robust techniques. Recently, some authors applied HD estimate approach to regression models. For example, Silvapullé [25] established asymptotic normality of HD estimators for the linear regression model with i.i.d. errors. Hu [1] investigated asymptotic normality of HD estimators for the linear regression model with AR(1) errors. Tong et al. [33] considered consistency and normality of HD estimators for the partial linear regression model. However, nobody used the HD method to investigate the models (1.1)-(1.2).
The paper discusses the models (1.1)-(1.2) with a robust approach, which has been suggested by Huber and Dutter. We extend some results of Hu [1], Silvapullé [25], etc. to the EV regression model with AR(1) errors. The organization of the paper is as follows. In Section 2 estimators of β, a and are given by HD method. Under general conditions, the asymptotic normality of the HD estimators is investigated in Section 3. The theoretical proofs of main results are presented in Section 4, a simple example is given in Section 5.
2 Estimation method
By (1.2), we have
thus is measurable with respect to the σ-field H generated by , and
Furthermore,
Let , . From (1.1),
By the above equation and (1.2), we obtain
Thus we could consider HD estimators by minimizing
where is convex, , as for some , and is a suitably chosen sequence of constants.
Remark 1 Since
using the method of HD, we also obtain HD estimators by minimizing
We will investigate its estimators in the future because there are some difficulties. For example, there is very sophisticated calculation, and it is difficult to investigate the asymptotic properties of these unknown parametric estimators because
are dependent.
Let us introduce some notation: vector and its estimator . For an arbitrary function f, and are the first and second derivatives of f, respectively. is the Euclidean norm of x and . and are the th component of matrix M and the i th component of vector U, respectively.
The HD estimators for θ are obtained by solving the estimating equations by equating to 0 the derivatives
and
where and .
The corresponding estimators, if they exist (see Proposition 2.1), will satisfy
and
with .
Although are unknown in (2.9)-(2.11), but we easily estimate it by the method of Fuller [17] in practice.
In what follows, it will be assumed that and . Without loss of generality, we may assume . (Its definition has been given between (2.5) and Remark 1.) Therefore ψ is bounded and increases from −1 to +1. It will also be assumed that χ is bounded.
Remark 2 From the above equations, it is easily seen that our estimators include some existing estimators; see, for example, simultaneous M-estimators of the location and the scale (, ), the least-squares estimators (, , and ), the least absolute deviation estimators (, ). In particular, we discuss three important cases as follows.
Case 1. Let , . The estimating equations (2.9)-(2.11) may be written as
which are the same as Silvapullé’s [25].
Case 2. If (), , and , then the above equations (2.9)-(2.11) may be rewritten as
and
with .
Let and (), . We rewrite (2.13)-(2.15) by
Furthermore, if σ is a constant, then the estimator of β satisfies the following equation:
which is or the maximum likelihood estimate equation for the parameter β in a linear regression model with q-norm distribution errors. Many authors have investigated (2.17), such as Arcones [34] and Zeckhauser and Thompson [35], Ronner [36, 37], and so on.
Case 3. Let . The estimating equations (2.9)-(2.11) may be written as
and
with , which are the same as Hu’s [1].
From Proposition 1 in Silvapullé [25] and pp.136 in Huber and Ronchetti [27], the existence results of the HD estimators may be given by the following.
Proposition 2.1 Suppose that is continuous and, for some , , where ν is the largest jump in the error distribution, , then the equation has a solution with . Especially, when , where is defined in (2.5), we denote it by with .
3 Main results
To obtain our results, we start with some assumptions.
-
(A1) .
-
(A2) for some , .
-
(A3) The function is continuous.
-
(A4) for any . , , , , , , , and .
-
(A5) For any , is positive definite for sufficiently large n.
Remark 3 The condition (A1) is often imposed in the estimation theory of regression models. The condition (A2) is used by Tong et al. [33]. In addition, by (A1) and (A2), we can obtain conditions and , which are used by Silvapullé [25]. The conditions (A3) and (A4) except and are used by Silvapullé [25]. The condition (A5) is used by Maller [2], Hu [1], etc. Therefore, our conditions are quite mild and can easily be satisfied.
For ease of exposition, we shall introduce the following notations which will be used later in the paper. Define
and
where the ∗ indicates that the elements are filled in by symmetry, and
and
Theorem 3.1 Suppose that conditions (A1)-(A5) hold. Then, as :
-
(1)
For and , we have
(3.9)
where and defined in Lemma 4.1.
-
(2)
For and , (3.9) holds.
-
(3)
For and , (3.9) holds.
From the above theorem, we may obtain the following corollaries. Here we omit their proofs.
Corollary 3.1 If and conditions (A2)-(A5) hold, then
and
Corollary 3.2 If and conditions (A1)-(A5) hold, then
Remark 4 Corollary 3.2 is similar to Theorem 2 of Silvapullé [25].
Corollary 3.3 Let σ be a constant. If conditions (A1)-(A5) hold, then
and
Corollary 3.4 Let . If conditions (A1)-(A5) hold, then Theorem 3.1 holds.
Remark 5 For , Corollary 3.4 is the same as Theorem 3.1 of Hu [1]. Therefore, we extend the corresponding results of Hu [1] to linear EV models.
If we do not consider the dependency on the parameters and , then we will obtain Theorem 3.2.
Theorem 3.2 Let
Suppose that conditions (A1)-(A5) hold. Then
-
(1)
for any ,
(3.15)
where
-
(2)
for any ,
(3.16)
4 Proofs of main results
Throughout this paper, let C denote a generic positive constant which could take different value at each occurrence. To prove Theorem 3.1 and Theorem 3.2, we first introduce the following lemmas.
Lemma 4.1 If (A4) and (A5) hold, then the matrix is positive definite with for sufficiently large n and
where
Furthermore, is a positive definite matrix.
Proof Note that , , and , , and are independent. By (3.3)-(3.8), we easily obtain
It is easy to show that
and
Thus the matrix is positive definite.
By (2.6), we have
By (2.7) and Proposition 2.1, we have
Note that and are independent; by (2.8) and , we have
Hence, from (4.4)-(4.6),
By (2.6), we have
By (2.7), we have
Note that is a martingale difference sequence with
so we have
By (2.6) and (2.7), we have
By (2.6), (2.8), and noting that , , and are independent, we have
By (2.7) and (2.8), we have
Hence, (4.1) follows immediately from (4.7)-(4.12). □
Similarly to the proof of , we easily prove that the matrix is positive definite. Thus, we complete the proof of Lemma 4.1.
Lemma 4.2 Assume that (A1) and (A4) hold. Then:
-
(1)
for , we have
(4.13) -
(2)
for , we have
(4.14)
where the ∗ indicates that the elements are filled in by symmetry;
-
(3)
for , we have
(4.15)
Proof By (3.3) and (4.2), we obtain
Note that are i.i.d. random variables with finite variance , we have
By the Chebyshev inequality and (4.17), we have
Similarly, we obtain
By (4.16), (4.18), and (4.19), we have
By (2.12) and (4.2), we easily obtain
By (3.5) and (4.2), we obtain
Note that is a martingale difference sequence with
so we have
By the Chebyshev inequality and (4.23), we have
Similarly, we have
It is easy to show that
By (4.22) and (4.24)-(4.26), we have
By (3.6) and (4.2), we obtain
Note that is a martingale difference sequence with , and by (3.7) and (4.2), we obtain
By (3.8) and (4.2), we obtain
Since we easily prove that is a martingale difference sequence,
By (2.1), we obtain
Thus
That is,
By Chebyshev inequality and (4.31)-(4.34), we have
Similarly to the proof of (4.35), we easily obtain
Hence, by (4.30), (4.35), and (4.36), we have
Thus Lemma 4.2 follows from (4.20), (4.21), (4.27)-(4.29), and (4.37). □
Lemma 4.3 Assume that (A1), (A3), and (A4) hold, and , and . Then, as :
-
(1)
for , we have
-
(2)
for , we have
-
(3)
for , we have
Proof Let
Case 1. .
and
Case 2. .
and
Case 3. .
and
□
Similarly to the proof of Lemma 4.2, we easily obtain Lemma 4.3. Here we omit it.
Lemma 4.4 (Prakasa Rao [38])
If are independent random variables with , and for some , then
where .
Lemma 4.5 (Hall and Heyde [39])
Let be a zero-mean, square-integrable martingale array with differences , and let be an a.s. finite random variable. Suppose that , for all , and . Then
where the r.v. Z has characteristic function .
Proof of Theorem 3.1 Expanding about θ, we have
where for some and
By (2.9)-(2.11) and (3.1), (3.2), we have
By (4.58), we have
(1) . By Lemma 4.3(1), Lemma 4.2(1), (2.9)-(2.11), and , we have
Note that , so we have
By Lemma 2 of Silvapullé [25], we easily obtain
In the following, we will prove that
Note that is a martingale differences sequence, so we will verify the Lindeberg conditions for their convergence to normality.
From (1.2), we have
By and Chebyshev inequality, we have
Obviously, . By (4.46) and , we have
By (4.65), we have
For given , there is a set whose probability approaches 1 as on which . In this event, for any ,
Here as . This verifies the Lindeberg conditions, hence (4.63) follows from Lemma 4.5.
Note that are asymptotic independent of
Therefore, we obtain Theorem 3.1(1) from (4.61)-(4.63).
(2) For . By Lemma 4.5(2), Lemma 4.2(2), (2.9)-(2.11), (4.59), and , we have
Note that , so we have
Similarly to the proof of (4.67) and (4.68), we have
and
This verifies the Lindeberg conditions, hence (4.63) follows from Lemma 4.5. Similarly to the proof of Theorem 3.1(1), we easily prove Theorem 3.1(2).
-
(3)
For . By Lemma 4.3(3), Lemma 4.2(3), (2.9)-(2.11), (4.59), and , we have
(4.73)
Note that , so we have
Similarly to the proof of Theorem 3.1(1), we easily prove Theorem 3.1(3). The proof of Theorem 3.1 is now complete. □
Proof of Theorem 3.2 Case 1. For . A first step towards (4.74) is to show that
Let with . Then
Thus,
By Lemma 4.4 and the Cramer-Wold device, (4.75) follows from (4.77).
Next we need to show that
In fact,
By Lemma 4.4, (4.78) follows from (4.79).
Finally, by (4.63), we easily prove that
Case 2. For . Similarly to the proof of Case 1, we easily prove this case.
This completes the proof of Theorem 3.2. □
5 Numerical example
In the section, we will simulate a simple linear regression model (1.1) with (1.2), where , , , and .
We take , , , , where and are the distribution and density function of standard normal , respectively Hu [1]. In the following, we calculate by using our method and the quasi-Newton line search method.
Case 1. For , we have , , and . and approximately equal β and σ, respectively.
Case 2. For , we have , and . For , we have , , and . and approximately equal β and a, respectively.
Case 3. For , we have , , and . and approximately equal β and a, respectively.
The above results show that our estimation method is valid in some cases.
References
Hu HH: Asymptotic normality of Huber-Dutter estimators in a linear model with AR(1) processes. J. Stat. Plan. Inference 2013,143(3):548–562. 10.1016/j.jspi.2012.08.012
Maller RA: Asymptotics of regressions with stationary and nonstationary residuals. 105. Stochastic Processes and Their Appliations 2003, 33–67.
Pere P: Adjusted estimates and Wald statistics for the AR(1) model with constant. J. Econom. 2000, 98: 335–363. 10.1016/S0304-4076(00)00023-3
Fuller WA: Introduction to Statistical Time Series. 2nd edition. Wiley, New York; 1996.
Miao Y, Liu W: Moderate deviations for LS estimator in simple linear EV regression model. J. Stat. Plan. Inference 2009,139(9):3122–3131. 10.1016/j.jspi.2009.02.021
Miao Y, Wang K, Zhao F: Some limit behaviors for the LS estimator in simple linear EV regression models. Stat. Probab. Lett. 2011, 81: 92–102. 10.1016/j.spl.2010.09.023
Miao Y, Yang G, Shen L: The central limit theorem for LS estimator in simple linear EV regression models. Commun. Stat., Theory Methods 2007, 36: 2263–2272. 10.1080/03610920701215266
Liu JX, Chen XR: Consistency of LS estimator in simple linear EV regression models. Acta Math. Sci. 2005, 25: 50–58.
Cui HJ: Asymptotic normality of M -estimates in the EV model. J. Syst. Sci. Math. Sci. 1997,10(3):225–236.
Cui HJ, Chen SX: Empirical likelihood confidence region for parameter in the errors-in-variables models. J. Multivar. Anal. 2003, 84: 101–115. 10.1016/S0047-259X(02)00017-9
Cheng CL, Van Ness JW: Statistical Regression with Measurement Error. Arnold, London; 1999.
Hamilton JD: Time Series Analysis. Princeton University Press, Princeton; 1994.
Brockwell PJ, Davis RA: Time Series: Theory and Methods. Springer, New York; 1987.
Baran S: A consistent estimator for linear models with dependent observations. Commun. Stat., Theory Methods 2004,33(10):2469–2486.
Fan GL, Liang HY, Wang JF, Xu HX: Asymptotic properties for LS estimators in EV regression model with dependent errors. AStA Adv. Stat. Anal. 2010, 94: 89–103. 10.1007/s10182-010-0124-3
Miao Y, Zhao F, Wang K: Central limit theorems for LS estimators in the EV regression model with dependent measurements. J. Korean Stat. Soc. 2011, 40: 303–312. 10.1016/j.jkss.2010.12.002
Fuller WA: Measurement Error Models. Wiley, New York; 1987.
Brown M: Robust line estimation with errors in both variables. J. Am. Stat. Assoc. 1982, 77: 71–79. 10.1080/01621459.1982.10477768
Ketellapper RH, Ronner AE: Are robust estimation methods useful in the structural errors-in-variables model. Metrika 1984, 31: 33–41. 10.1007/BF01915180
Zamar RH: Robust estimation in the errors-in-variables model. Biometrika 1989,76(1):149–160. 10.1093/biomet/76.1.149
Cheng CL, Van Ness JW: Generalized M -estimators for errors-in-variables regression. Ann. Stat. 1992,20(1):385–397. 10.1214/aos/1176348528
He X, Liang H: Quantile regression estimates for a class of linear and partially linear errors-in-variables models. Stat. Sin. 2000, 10: 129–140.
Fekri M, Ruiz-Gazen A: Robust weighted orthogonal regression in the errors-in-variables model. J. Multivar. Anal. 2004, 88: 89–108. 10.1016/S0047-259X(03)00057-5
Wu WB: M -Estimation of linear models with dependent errors. Ann. Stat. 2007,35(2):495–521. 10.1214/009053606000001406
Silvapullé MJ: Asymptotic behavior of robust estimators of regression and scale parameters with fixed carriers. Ann. Stat. 1985,13(4):1490–1497. 10.1214/aos/1176349750
Hampel FR, Ronchetti EM, Rousseeuw PJ, et al.: Robust Statistics. Wiley, New York; 1986.
Huber PJ, Ronchetti EM: Robust Statistics. 2nd edition. Wiley, New York; 2009.
Li L: On Koul’s minimum distance estimators in the regression models with long memory moving averages. 105. Stochastic Processes and Their Applications 2003, 257–269.
Babu GJ: Strong representations for LAD estimators in linear models. Probab. Theory Relat. Fields 1989, 83: 547–558. 10.1007/BF01845702
Salibian-Barrera M, Zamar RH: Bootstrapping robust estimates of regression. Ann. Stat. 2002,30(2):556–582. 10.1214/aos/1021379865
Wu R, Wang Q: Shrinkage estimation for linear regression with ARMA errors. J. Stat. Plan. Inference 2012, 142: 2136–2148. 10.1016/j.jspi.2012.02.047
Zhou Z, Wu W: On linear models with long memory and heavy-tailed errors. J. Multivar. Anal. 2011, 102: 349–362. 10.1016/j.jmva.2010.09.009
Tong XW, Cui HJ, Yu P: Consistency and normality of Huber-Dutter estimators for partial linear model. Sci. China Ser. A 2008,51(10):1831–1842. 10.1007/s11425-008-0028-9
Arcones MA:The Bahadur-Kiefer representation of regression estimators. Econom. Theory 1996, 12: 257–283. 10.1017/S0266466600006587
Zeckhauser R, Thompson M: Linear regression with non-normal error terms. Rev. Econ. Stat. 1970, 52: 280–286. 10.2307/1926296
Ronner, AE: P-norm estimators in a linear regression model. Ph.D. thesis, Drukkerijenbv-Groningen, The Netherlands (1977)
Ronner AE: Asymptotic normality of p -norm estimators in multiple regression. Z. Wahrscheinlichkeitstheor. Verw. Geb. 1984, 66: 613–620. 10.1007/BF00531893
Prakasa Rao BLS: Asymptotic Theory of Statistical Inference. Wiley, New York; 1987.
Hall P, Heyde CC: Martingale Limit Theory and Its Application. Academic Press, New York; 1980.
Acknowledgements
The first author’s work was supported by the Natural Science Foundation of China (No. 11471105). The second author’s work was Supported by Natural Science Foundation of China (No. 41374017).
Author information
Authors and Affiliations
Corresponding author
Additional information
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
All authors contributed equally to the writing of this paper. All authors read and approved the final manuscript.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0), which permits use, duplication, adaptation, distribution, and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
About this article
Cite this article
Hu, H., Pan, X. Asymptotic normality of Huber-Dutter estimators in a linear EV model with AR(1) processes. J Inequal Appl 2014, 474 (2014). https://doi.org/10.1186/1029-242X-2014-474
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/1029-242X-2014-474