Abstract
We use distance covariance to introduce novel consistent tests of heteroscedasticity for nonlinear regression models in multidimensional spaces. The proposed tests require no user-defined regularization, which are simple to implement based on only pairwise distances between points in the sample and are applicable even if we have non-normal errors and many covariates in the regression model. We establish the asymptotic distributions of the proposed test statistics under the null and alternative hypotheses and a sequence of local alternatives converging to the null at the fastest possible parametric rate. In particular, we focus on whether and how the estimation of the finite-dimensional unknown parameter vector in regression functions will affect the distribution theory. It turns out that the asymptotic null distributions of the suggested test statistics depend on the data generating process, and then a bootstrap scheme and its validity are considered. Simulation studies demonstrate the versatility of our tests in comparison with the score test, the Cramér-von Mises test, the Kolmogorov-Smirnov test and the Zheng-type test. We also use the ultrasonic reference block data set from National Institute for Standards and Technology of USA to illustrate the practicability of our proposals.
Similar content being viewed by others
References
Biswas M, Mukhopadhyay M, Ghosh A K. A distribution-free two-sample run test applicable to high-dimensional data. Biometrika, 2014, 101: 913–926
Chakraborty S, Zhang X Y. Distance metrics for measuring joint dependence with application to causal inference. J Amer Statist Assoc, 2019, 114: 1638–1650
Cheng H, Li Q. A consistent test for conditional heteroskedasticity in time-series regression models. Econometric Theory, 2001, 17: 188–221
Cook R D, Weisberg S. Diagnostics for heteroscedasticity in regression. Biometrika, 1983, 70: 1–10
Escanciano J C. A consistent diagnostic test for regression models using projections. Econometric Theory, 2006, 22: 1030–1051
Fokianos K, Pitsillou M. Testing independence for multivariate time series via the auto-distance correlation matrix. Biometrika, 2018, 105: 337–352
Gretton A, Bousquet O, Smola A J, et al. Measuring statistical dependence with Hilbert-Schmidt norms. In: International Conference on Algorithmic Learning Theory. Lecture Notes in Artificial Intelligence, vol. 3734. Berlin: Springer, 2005, 63–77
Guo X, Jiang X J, Zhang S M, et al. Pairwise distance-based heteroscedasticity test for regressions. Sci China Math, 2020, 63: 2553–2572
Heuchenne C, Van Keilegom I. Nonlinear regression with censored data. Technometrics, 2007, 49: 34–44
Huet S, Bouvier A, Poursat M A, et al. Statistical Tools for Nonlinear Regression. New York: Springer-Verlag, 2004
Huo X M, Székely G J. Fast computing for distance covariance. Technometrics, 2016, 58: 435–447
Jennrich R I. Asymptotic properties of nonlinear least squares estimators. Ann Math Statist, 1969, 40: 633–643
Knight K. Limiting distributions for L1 regression estimators under general conditions. Ann Statist, 1998, 26: 755–770
Lee C E, Shao X F. Martingale difference divergence matrix and its application to dimension reduction for stationary multivariate time series. J Amer Statist Assoc, 2018, 113: 216–229
Lee Y J, Shen C C, Priebe C E, et al. Network dependence testing via diffusion maps and distance-based correlations. Biometrika, 2019, 106: 857–873
Leucht A, Neumann M H. Consistency of general bootstrap methods for degenerate U-type and V-type statistics. J Multivariate Anal, 2009, 100: 1622–1633
Li R Z, Zhong W, Zhu L P. Feature screening via distance correlation learning. J Amer Statist Assoc, 2012, 107: 1129–1139
Matteson D S, Tsay R S. Independent component analysis via distance covariance. J Amer Statist Assoc, 2017, 112: 623–637
Pakes A, Pollard D. Simulation and the asymptotics of optimization estimators. Econometrica, 1989, 57: 1027–1057
Pan W L, Wang X Q, Zhang H P, et al. Ball covariance: A generic measure of dependence in Banach space. J Amer Statist Assoc, 2020, 115: 307–317
Patra R K, Sen B, Székely G J. On a nonparametric notion of residual and its applications. Statist Probab Lett, 2016, 109: 208–213
Sen A, Sen B. Testing independence and goodness-of-fit in linear models. Biometrika, 2014, 101: 927–942
Sen P K. Estimates of the regression coefficient based on Kendall’s tau. J Amer Statist Assoc, 1968, 63: 1379–1389
Serfling R L. Approximation Theorems in Mathematical Statistics. New York: Wiley, 1980
Shao X F, Zhang J S. Martingale difference correlation and its use in high-dimensional variable screening. J Amer Statist Assoc, 2014, 109: 1302–1318
Shen C C, Priebe C E, Vogelstein J T. From distance correlation to multiscale graph correlation. J Amer Statist Assoc, 2020, 115: 280–291
Sheng W H, Yin X R. Sufficient dimension reduction via distance covariance. J Comput Graph Statist, 2016, 25: 91–104
Sherman R P. Maximal inequalities for degenerate U-processes with applications to optimization estimators. Ann Statist, 1994, 22: 439–459
Stute W. Nonparametric model checks for regression. Ann Statist, 1997, 25: 613–641
Stute W, Xu W L, Zhu L X. Model diagnosis for parametric regression in high-dimensional spaces. Biometrika, 2008, 95: 451–467
Su L J, Ullah A. A nonparametric goodness-of-fit-based test for conditional heteroskedasticity. Econometric Theory, 2013, 29: 187–212
Székely G J, Rizzo M L. Brownian distance covariance. Ann Appl Stat, 2009, 3: 1236–1265
Székely G J, Rizzo M L. Partial distance correlation with methods for dissimilarities. Ann Statist, 2014, 42: 2382–2412
Székely G J, Rizzo M L, Bakirov N K. Measuring and testing dependence by correlation of distances. Ann Statist, 2007, 35: 2769–2794
Tan F L, Jiang X J, Guo X, et al. Testing heteroscedasticity for regression models based on projections. Statist Sinica, 2021, 31: 625–646
Van der Vaart A W. Asymptotic Statistics. Cambridge: Cambridge University Press, 2000
Van Keilegom I, González-Manteiga W, Sánchez-Sellero C. Goodness-of-fit tests in parametric regression based on the estimation of the error distribution. TEST, 2008, 17: 401–415
Wang L, Zhou X H. Assessing the adequacy of variance function in heteroscedastic regression models. Biometrics, 2007, 63: 1218–1225
Wang X Q, Pan W L, Hu W H, et al. Conditional distance correlation. J Amer Statist Assoc, 2015, 110: 1726–1734
Xu K, He D. Omnibus model checks of linear assumptions through distance covariance. Statist Sinica, 2021, 31: 1055–1079
Yao S, Zhang X Y, Shao X F. Testing mutual independence in high dimension via distance covariance. J R Stat Soc Ser B Stat Methodol, 2018, 80: 455–480
Zheng J X. A consistent test of functional form via nonparametric estimation techniques. J Econometrics, 1996, 75: 263–289
Zheng J X. Testing heteroscedasticity in nonlinear and nonparametric regressions. Canad J Statist, 2009, 37: 282–300
Zhu L P, Xu K, Li R Z, et al. Projection correlation between two random vectors. Biometrika, 2017, 104: 829–843
Zhu L X, Fujikoshi Y, Naito K. Heteroscedasticity checks for regression models. Sci China Ser A, 2001, 44: 1236–1252
Acknowledgements
This work was supported by National Natural Science Foundation of China (Grant Nos. 11901006 and 11601008) and Natural Science Foundation of Anhui Province (Grant No. 1908085QA06).
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Xu, K., Cao, M. Distance-covariance-based tests for heteroscedasticity in nonlinear regressions. Sci. China Math. 64, 2327–2356 (2021). https://doi.org/10.1007/s11425-020-1759-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11425-020-1759-5