Abstract
In this paper we address the issue of testing inference of the dispersion parameter in heteroscedastic symmetric nonlinear regression models considering small samples. We derive Bartlett corrections to improve the likelihood ratio as well modified profile likelihood ratio tests. Our results extend some of those obtained in Cordeiro (J Stat Comput Simul 74:609–620, 2004) and Ferrari et al. (J Stat Plan Inference 124:423–437, 2004), who consider a symmetric nonlinear regression model and normal linear regression model, respectively. We also present the bootstrap and bootstrap Bartlett corrected likelihood ratio tests. Monte Carlo simulations are carried out to compare the finite sample performances of the three corrected tests and their uncorrected versions. The numerical evidence shows that the corrected modified profile likelihood ratio test, the bootstrap and bootstrap Bartlett corrected likelihood ratio test perform better than the other ones. We also present an empirical application.
Similar content being viewed by others
Notes
Let \(\{a_n\}\) and \(\{ b_n\}\) be real sequences. \(a_n\) is said to be at most of order equal to \(b_n,\) denoted by \(a_n=O(b_n),\) if \(K \in {\mathbb {R}}^+\) exists and \(n_0(K),\) such that \(|a_n/b_n|\le K, \ \forall n\ge n_0(K).\)
References
Bartlett MS (1937) Properties of sufficiency and statistical tests. Proc R Soc Lond A 160:268–282
Bayer FM, Cribari-Neto F (2013) Bartlett corrections in beta regression models. J Stat Plan Inference 143:531–547
Cook D, Weisberg S (1983) Diagnostics for heteroscedasticity in regression. Biometrika 70:1–10
Cordeiro GM (2004) Corrected likelihood ratio tests in symmetric nonlinear regression models. J Stat Comput Simul 74:609–620
Cordeiro GM, Cribari-Neto F (2014) An introduction to Bartlett correction and bias correction. Springer, New York
Cordeiro GM, Ferrari SLP, Uribe-Opazo MA, Vasconcellos KLP (2000) Corrected maximum-likelihood estimation in a class of symmeytric nonlinear regression models. Stat Probab Lett 46:317–328
Cox DR, Reid N (1987) Parameter orthogonality and approximate conditional inference. J R Stat Soc B 49:1–39
Cox DR, Reid N (1992) A note on the difference between profile and modified profile likelihood. Biometrika 79:408–411
Cribari-Neto F, Queiroz MPF (2014) On testing inference in beta regression. J Stat Comput Simul 84:186–203
Cysneiros FJA, Cordeiro GM, Cysneiros AHMA (2010a) Corrected maximum likelihood estimators in heteroscedastic symmetric nonlinear models. J Stat Comput Simul 80:451–461
Cysneiros AHMA, Ferrari SLP (2006) An improved likelihood ratio test for varying dispersion in exponential family nonlinear models. Stat Prob Lett 76:255–265
Cysneiros FJA, Paula GA, Galea M (2005) Modelos Simétricos Aplicados. 9 Escola de Modelos de Regressão, São Paulo
Cysneiros AHMA, Rodrigues KSP, Cordeiro GM, Ferrari SLP (2010b) Three Bartlett-type corrections for score statistics in symmetric nonlinear regression models. Stat Pap 51:273–284
DiCiccio TJ, Stern SE (1994) Frequentist and Bayesian Bartlett correction of test statistics based on adjusted profile likelihoods. J R Stat Soc Ser B Methodol 56:397–408
Doornik JA (2006) An object-oriented matrix language Ox 5. Timberlake Consultants Press, London
Efron B (1979) Bootstrap methods: another look at the jackknife. Ann Stat 7:1–26
Fang KT, Anderson TW (1990b) Statistical inference in elliptical contoured and related distribuitions. Allerton Press, New York
Fang KT, Kotz S, Ng KW (1990a) Symmetric multivariate and related distributions. Chapman and Hall, London
Ferrari SLP, Cysneiros AHMA, Cribari-Neto F (2004) An improved test for heteroskedasticity using adjusted modified profile likelihood inference. J Stat Plan Inference 124:423–437
Ferrari SLP, Lucambio F, Cribari-Neto F (2005) Improved profile likelihood inference. J Stat Plan Inference 134:373–391
Fujita A, Kojima K, Patriota AG, Sato JR, Severino P, Miyano S (2010) A fast and robust test based on likelihood ratio with Bartlett correction to identify granger causality between gene sets. Bioinformatics 26:2349–2351
Lawley DN (1956) A general method for approximating to the distribution of the likelihood ratio criteria. Biometrika 71:233–244
Lemonte AJ, Cordeiro GM, Moreno G (2012) Bartlett corrections in Birnbaum-Saunders nonlinear regression models. J Stat Comput Simul 82:927–935
Lin JG, Zhu LX, Xie FG (2009) Heteroscasticity diagnostics for \(t\) linear regression models. Metrika 70:59–77
Loose LH, Bayer FM, Liberal TL (2016) Bootstrap Bartlett corretion in inflated beta regression. Commun Stat Simul 46(4):2865–2879. doi:10.1080/03610918.2015.1065326
Maior VQS, Cysneiros FJA (2016) SYMARMA: a new dynamic model for temporal data on conditional symmetric distribution. Stat Pap. doi:10.1007/s00362-016-0753-z
Melo TFN, Ferrari SLP, Cribari-Neto F (2009) Improved test inference in mixed linear models. Comp Stat Data Anal 53:2573–2582
Nocedal J, Wright SJ (1999) Numerical optimization. Springer, New York
Pace L, Salvan A (1997) Principles of statistical inference from a Neo-Fisherian perspective. World Scientific, Singapore
Rocke DM (1989) Bootstrap Bartlett adjustment in seemingly unrelated regression. J Am Stat Assoc 84:598–601
Simonoff JS, Tsai CH (1994) Use of modified profile likelihood for improved tests of constancy of variance in regression. Appl Stat 43:357–370
Stein MC, Silva MF, Duczmal LH (2014) Alternatives to the usual likelihood ratio test in mixed linear models. Comp Stat Data Anal 69:184–197
Uribe-Opazo MA, Ferrari SLP, Cordeiro GM (2008) Improved score test in symmetric linear regression model. Commun Stat Theory Methods 37:261–276
Vanegas LH, Rondon LM, Cysneiros FJA (2013) Assessing robustness of inference in symmetrical nonlinear regression models. Commun Stat Theory Methods 42:1692–1711
Wei BL (1998) Exponential family nonlinear models. Springer, Singapore
Acknowledgements
We would like to thank the reviewers for their constructive comments, which helped to improve this manuscript. Also, we gratefully acknowledge the financial support of CAPES, CNPq, FAPEMIG and FACEPE.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix A
In this appendix we obtain the matrix \(j^*({\varvec{\delta }}; \varvec{\hat{\beta }_\delta }, \hat{\gamma }_\delta )\) expressed in (3) which denotes the observed information matrix for the nuisance parameters \(({\varvec{\beta }}^\top , \gamma )^\top . \) We can express \(j^*({\varvec{\delta }}; \varvec{\hat{\beta }_\delta }, \hat{\gamma }_\delta )\) by
where \(j^{*}_{\beta \beta }\) is a square matrix of order p whose entries are given by
and
is a scalar, with \(\varvec{\theta ^*}=({\varvec{\beta }}^\top , {\varvec{\delta }}^\top ,\gamma )^\top \,\,q_\ell =\frac{(\prod _{s=1}^n m_s)^{1/n}}{m_\ell }\), \((j,l)_\ell = (\partial \mu _\ell /\partial \beta _j)\,\,(\partial \mu _\ell / \partial \beta _l)\) and \((jl)_\ell =\partial ^2 \mu _\ell / \partial \beta _j \partial \beta _l, \ j,l=1,\ldots ,p.\)
Appendix B
In this appendix we present the required derivatives of the log-likelihood function in (2) up to the fourth order and their respective moments, considering \(m_\ell =\exp (\varvec{\omega }_\ell ^\top {\varvec{\delta }}),\) to obtain the constants c and \(c_m\) from the Bartlett correction factors. For this, we introduce the following notation: \(\lambda _{rs}=E(\partial ^2 l/\partial {\varvec{\theta }}_r \partial {\varvec{\theta }}_s)\), \(\lambda _{rst}=E(\partial ^3 l/\partial {\varvec{\theta }}_r \partial {\varvec{\theta }}_s \partial {\varvec{\theta }}_t),\) etc, \(a, b, c,\ldots \) index parameters of interest \({\varvec{\delta }}\), i, j, \(l,\ldots \) index nuisance parameters \({\varvec{\beta }}\) and \(r,s,t,\ldots \) index all \(p+q+1\) parameters of the model. Thus, we have
and the derived cumulants required to calculate c and \(c_m\) are given by:
Appendix C
Here we obtain in detail the expression c for the Bartlett correction factor for the LR statistic. As defined in Sect. 3.1, the Bartlett correction factor is expressed by \(1+c/k,\) where
From Lawley’s expansion (1956), we have that \(E(LR)=k+\epsilon _{p,k}-\epsilon _p+O(n^{-2}),\) where
being
where \(-\lambda ^{rs}\) is the (r, s) element from the inverse of Fisher’s information matrix. The sum in (6) varies in all components of \({\varvec{\theta }}^*\). The expression \(\epsilon _p\) is obtained from (6) considering that the sum varies across all nuisance parameters. Thus, we can write c as
where \(\epsilon _k ({\varvec{\delta }})\) indicates that the sum (6) was performed on all components of \({\varvec{\delta }}\), analogously to other cases, considering the respective parameters. Replacing \(\lambda '\)s obtained in Appendix B in the \(\epsilon '\)s and after of extensive algebra, we have the constant c in (4).
Appendix D
Finally, here we obtain the constant \(c_m\) for the Bartlett correction factor for the \(LR_m\) statistic, considering Equation (5) of Ferrari et al. (2004). Thus, we have that \(c_m\) can be expressed by
Replacing the cumulant \(\lambda '\)s obtained in Appendix B in Equation (7) and after intense algebra, we have
where \((\varvec{\omega }_\ell - \bar{\varvec{\omega }})_i = \omega _{\ell i}-\bar{\omega }_i,\) with \(i=a,b,c,d.\) In matrix notation, we have
Rights and permissions
About this article
Cite this article
Araújo, M.C., Cysneiros, A.H.M.A. & Montenegro, L.C. Improved heteroskedasticity likelihood ratio tests in symmetric nonlinear regression models. Stat Papers 61, 167–188 (2020). https://doi.org/10.1007/s00362-017-0933-5
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00362-017-0933-5