Electronic Journal of Statistics

Minimax testing of a composite null hypothesis defined via a quadratic functional in the model of regression

Laëtitia Comminges and Arnak S. Dalalyan

Full-text: Open access


We consider the problem of testing a particular type of composite null hypothesis under a nonparametric multivariate regression model. For a given quadratic functional $Q$, the null hypothesis states that the regression function $f$ satisfies the constraint $Q[f]=0$, while the alternative corresponds to the functions for which $Q[f]$ is bounded away from zero. On the one hand, we provide minimax rates of testing and the exact separation constants, along with a sharp-optimal testing procedure, for diagonal and nonnegative quadratic functionals. We consider smoothness classes of ellipsoidal form and check that our conditions are fulfilled in the particular case of ellipsoids corresponding to anisotropic Sobolev classes. In this case, we present a closed form of the minimax rate and the separation constant. On the other hand, minimax rates for quadratic functionals which are neither positive nor negative makes appear two different regimes: “regular” and “irregular”. In the “regular" case, the minimax rate is equal to $n^{-1/4}$ while in the “irregular” case, the rate depends on the smoothness class and is slower than in the “regular” case. We apply this to the problem of testing the equality of Sobolev norms of two functions observed in noisy environments.

Article information

Electron. J. Statist., Volume 7 (2013), 146-190.

First available in Project Euclid: 24 January 2013

Permanent link to this document

Digital Object Identifier

Mathematical Reviews number (MathSciNet)

Zentralblatt MATH identifier

Primary: 62G08: Nonparametric regression 62G10: Hypothesis testing
Secondary: 62G20: Asymptotic properties

Nonparametric hypotheses testing sharp asymptotics separation rates minimax approach high-dimensional regression


Comminges, Laëtitia; Dalalyan, Arnak S. Minimax testing of a composite null hypothesis defined via a quadratic functional in the model of regression. Electron. J. Statist. 7 (2013), 146--190. doi:10.1214/13-EJS766. https://projecteuclid.org/euclid.ejs/1359041588

Export citation


  • Baraud, Y., Huet, S., and Laurent, B. Adaptive tests of linear hypotheses by model selection., Ann. Statist., 31(1):225–251, 2003.
  • Baraud, Y., Huet, S., and Laurent, B. Testing convex hypotheses on the mean of a Gaussian vector. Application to testing qualitative hypotheses on a regression function., Ann. Statist., 33(1):214–257, 2005.
  • Butucea, C. Goodness-of-fit testing and quadratic functional estimation from indirect observations., Ann. Statist., 35(5) :1907–1930, 2007.
  • Butucea, C., Matias, C., and Pouet, C. Adaptive goodness-of-fit testing from indirect observations., Ann. Inst. Henri Poincaré Probab. Stat., 45 (2):352–372, 2009.
  • Cohen, A., Numerical analysis of wavelet methods, volume 32 of Studies in Mathematics and its Applications. North-Holland Publishing Co., Amsterdam, 2003.
  • Collier, O. Minimax hypothesis testing for curve registration., Electron. J. Statist., 6 :1129–1154, 2012.
  • Comminges, L. and Dalalyan, A. S. Tight conditions for consistent variable selection in high dimensional nonparametric regression., Journal of Machine Learning Research - Proceedings Track, 19:187–206, 2011.
  • Comminges, L. and Dalalyan, A. S. Tight conditions for consistency of variable selection in the context of high dimensionality., Ann. Statist., 40(6), doi:10.1214/12-AOS1046, 2012.
  • Dalalyan, A. S. and Collier, O. Wilks’ phenomenon and penalized likelihood-ratio test for nonparametric curve registration., J. Mach. Learn. Res. - Proceedings Track, 22:264–272, 2012.
  • Dalalyan, A. S., Juditsky, A., and Spokoiny, V. A new algorithm for estimating the effective dimension-reduction subspace., J. Mach. Learn. Res., 9 :1648–1678, 2008.
  • Donoho, D. and Nussbaum, M. Minimax quadratic estimation of a quadratic functional., J. Complexity, 6(3):290–323, 1990.
  • Efromovich, S. On the limit in the equivalence between heteroscedastic regression and filtering model., Statist. Probab. Lett., 63(3):239–242, 2003.
  • Ermakov, M. S. Minimax detection of a signal in Gaussian white noise., Teor. Veroyatnost. i Primenen., 35(4):704–715, 1990.
  • Ermakov, M. S. Minimax detection of a signal in weighted Gaussian white noise., Zap. Nauchn. Sem. S.-Peterburg. Otdel. Mat. Inst. Steklov. (POMI), 320(Veroyatn. i Stat. 8):54–68, 226, 2004.
  • Ermakov, M. S. Nonparametric signal detection with small type I and type II error probabilities., Stat. Inference Stoch. Process., 14(1):1–19, 2011.
  • Fan, J. On the estimation of quadratic functionals., Ann. Statist., 19(3) :1273–1294, 1991.
  • Gaïffas, S. and Lecué, G. Optimal rates and adaptation in the single-index model using aggregation., Electron. J. Stat., 1:538–573, 2007.
  • Gayraud, G. and Pouet, C. Minimax testing composite null hypotheses in the discrete regression scheme., Math. Methods Statist., 10(4):375–394 (2002), 2001. Meeting on Mathematical Statistics (Marseille, 2000).
  • Gayraud, G. and Pouet, C. Adaptive minimax testing in the discrete regression scheme., Probab. Theory Related Fields, 133(4):531–558, 2005.
  • Hall, P. Central limit theorem for integrated square error of multivariate nonparametric density estimators., J. Multivariate Anal., 14(1):1–16, 1984.
  • Horowitz, J. L. and Spokoiny, V. G. An adaptive, rate-optimal test of a parametric mean-regression model against a nonparametric alternative., Econometrica, 69(3):599–631, 2001.
  • Ingster, Yu. and Stepanova, N. Estimation and detection of functions from anisotropic Sobolev classes., Electron. J. Stat., 5:484–506, 2011.
  • Ingster, Yu. I. Asymptotically minimax hypothesis testing for nonparametric alternatives. I., Math. Methods Statist., 2(2):85–114, 1993a.
  • Ingster, Yu. I. Asymptotically minimax hypothesis testing for nonparametric alternatives. II., Math. Methods Statist., 2(3):171–189, 1993b.
  • Ingster, Yu. I. Asymptotically minimax hypothesis testing for nonparametric alternatives. III., Math. Methods Statist., 2(4):249–268, 1993c.
  • Ingster, Yu. I. and Sapatinas, T. Minimax goodness-of-fit testing in multivariate nonparametric regression., Math. Methods Statist., 18(3):241–269, 2009.
  • Ingster, Yu. I., Sapatinas, T., and Suslina, I. A. Minimax Signal Detection in Ill-Posed Inverse Problems., Ann. Statist., 40(2) :1524–1549, 2012.
  • Ingster, Yu. I. and Suslina, I. A., Nonparametric goodness-of-fit testing under Gaussian models, volume 169 of Lecture Notes in Statistics. Springer-Verlag, New York, 2003.
  • Kneser, H. Sur un théorème fondamental de la théorie des jeux., C. R. Acad. Sci. Paris, 234 :2418–2420, 1952.
  • Kolyada, V. I. On the embedding of Sobolev spaces., Mat. Zametki, 54(3):48–71, 158, 1993.
  • Laurent, B., Loubes, J. M., and Marteau, C. Testing inverse problems: a direct or an indirect problem?, Journal of Statistical Planning and Inference, 141 (5) :1849–1861, 2011.
  • Laurent, B., Loubes, J. M., and Marteau, C. Non asymptotic minimax rates of testing in signal detection with heterogeneous variances., Electronic Journal of Statistics, 6:91–122, 2012.
  • Lepski, O., Nemirovski, A., and Spokoiny, V. On estimation of the $L_r$ norm of a regression function., Probab. Theory Related Fields, 113(2):221–253, 1999.
  • Pinsker, M. S. Optimal filtration of square-integrable signals in Gaussian noise., Problems Inform. Transmission, (16):52–68, 1980.
  • Pouet, C. An asymptotically optimal test for a parametric set of regression functions against a non-parametric alternative., J. Statist. Plann. Inference, 98(1-2):177–189, 2001.
  • Samarov, A., Spokoiny, V., and Vial, C. Component identification and estimation in nonlinear high-dimensional regression models by structural adaptation., J. Amer. Statist. Assoc., 100(470):429–445, 2005.
  • Spokoiny, V. G. Adaptive hypothesis testing using wavelets., Ann. Statist., 24(6) :2477–2498, 1996.
  • Tsybakov, A. B., Introduction to nonparametric estimation. Springer Verlag, 2009.
  • Vershynin, R. Introduction to the non-asymptotic analysis of random matrices., Compressed sensing, 210–268, Cambridge Univ. Press, Cambridge, 2012.