Abstract
The paper deals with the asymptotic behaviour of estimators, statistical tests and confidence intervals for \(L^2\)-distances to uniformity based on the empirical distribution function, the integrated empirical distribution function and the integrated empirical survival function. Approximations of power functions, confidence intervals for the \(L^2\)-distances and statistical neighbourhood-of-uniformity validation tests are obtained as main applications. The finite sample behaviour of the procedures is illustrated by a simulation study.
Similar content being viewed by others
References
Abate J, Valkó PP (2004) Multi-precision Laplace transform inversion. Int J Numer Methods Eng 60:979–993
Achieser NI (1992) Theory of approximation. Dover Publications, New York
Bahadur RR (1971) Some limit theorems in statistics. SIAM, Philadelphia
Baringhaus L, Henze N (2017) Cramér-von Mises distance: probabilistic interpretation, confidence intervals, and neighbourhood-of-model validation. J Nonparametr Stat 29:167–188
Baringhaus L, Ebner B, Henze N (2017) The limit distribution of weighted \(L^2\)-goodness-of-fit statistics under fixed alternatives, with applications. Ann Inst Stat Math 69:969–995
Bingham N, Goldie C, Teugels J (1989) Regular variation. Cambridge University Press, Cambridge
Courant R, Hilbert D (1968) Methoden der Mathematischen Physik I. Springer, Berlin
Csörgő S, Faraway J (1996) The exact and asymptotic distributions of Cramér–von Mises statistics. J R Stat Soc Ser B 58:221–234
Durbin J, Knott M (1972) Components of Cramér–von Mises statistics. I. J R Stat Soc Ser B 34:290–307
Gao F, Hannig J, Lee T-Y, Torcaso F (2003) Laplace transforms via Hadamard factorization. Electron J Probab 8(13):20
Henze N, Nikitin Y (2000) A new approach to goodness-of-fit testing based on the integrated empirical process. J Nonparametr Stat 12:391–416
Kallenberg WCM, Ledwina T (1987) On local and nonlocal measures of efficiency. Ann Stat 15:1401–1420
Kamke E (1983) Differentialgleichungen. Lösungsmethoden und Lösungen I. Teubner, Stuttgart
Khoshnevisan D, Shi Z (1998) Chung’s law for integrated Brownian motion. Trans Am Math Soc 350:4253–4264
Klar B (2001) Goodness-of-fit test for the exponential and the normal distribution based on the integrated empirical distribution function. Ann Inst Stat Math 53:338–353
Kourouklis S (1989) On the relation between Hodges–Lehmann efficiency and pitman efficiency. Can J Stat 17:311–318
Nazarov A, Nikitin Y (2004) Exact \(L_2\)-small ball behavior of integrated Gaussian processes and spectral asymptotics of boundary value problems. Probab Theory Relat Fields 129:469–494
Nikitin Y (1995) Asymptotic efficiency of nonparametric tests. Cambridge University Press, Cambridge
Pollard D (1984) Convergence of stochastic processes. Springer, New York
Serfling R (1980) Approximation theorems of mathematical statistics. Wiley, New York
Titchmarsh EC (1939) The theory of functions. Oxford University Press, Oxford
Wellek S (2010) Testing statistical hypotheses of equivalence and noninferiority. CRC Press, Boca Raton
Acknowledgements
The authors thank the referees for constructive comments and suggestions.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
We introduce the entire function
and note that the real positive zeros of d are just the positive solutions \(\kappa _j,\,j\ge 1,\) of the equation (2.1). We aim to derive the representation of d as an infinite product involving the zeros of d. Obviously, 0 is a simple zero of d. Observing
and
for each \(z\in \mathbb {C},\) we note that if \(0\ne z_0\in \mathbb {C}\) is a zero of d, also \(-z_0,~iz_0,\) and \(-iz_0\) are zeros of d. Thus, with \(\kappa _k,k\in \mathbb {N},\) as the real positive zeros of d, we recognise that \(-\kappa _k,k\in \mathbb {N},\) \(i\kappa _k,k\in \mathbb {N},\) and \(-i\kappa _k,k\in \mathbb {N},\) are zeros of d. Adding the zero 0, we assert that this is the whole set of zeros of d. To see that the assertion is true, it suffices to prove the following lemma.
Lemma 4
There is no zero \(z=x+iy\) of d with real positive x and real positive y.
Proof
Assume, there is a zero \(z=x+iy\) of d with real positive x and real positive y. Then,
Due to \(\sin u +\sinh u>0\) for real positive u, it is \(x\ne y.\) We can (and do) assume without loss of generality that \(x>y.\) Putting \(a=x+y,~b=x-y\) we have
From this, we deduce that
Adding the corresponding terms on the right and the left hand side in (6.2) and using (6.2) we obtain
and
It follows that
This is impossible, because the function \(\cosh u + \cos u,\,u>0,\) is strictly increasing. \(\square \)
Theorem 4
(Gao et al. (2003)) The Laplace Transform of the limit distribution is given by
where \(z(t)=2^{1/2}(2t)^{1/4}\) for \(t\ge 0.\)
Proof
Note that
is an entire function of order 1 and that \(f(0)=1.\) Due to
the genus of the canonical product of the primary factors of f is 1. So, by definition, f is of genus 1. By Hadamard’s factorisation theorem, see, e.g., Titchmarsh (1939), Sec. 8.24, it follows that f has the representation
where \(Q(z)=\alpha z, z\in \mathbb {C},\) with \(\alpha \in \mathbb {C}.\) Since \(f(z)=f(-z)\) for \(z\in \mathbb {C}\) it is \(\alpha =0.\) Thus,
and the Laplace transform of
is seen to be
\(\square \)
Rights and permissions
About this article
Cite this article
Baringhaus, L., Gaigall, D. & Thiele, J.P. Statistical inference for \(L^2\)-distances to uniformity. Comput Stat 33, 1863–1896 (2018). https://doi.org/10.1007/s00180-018-0820-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00180-018-0820-0