Abstract
We study the detection of change-points in time series. The classical CUSUM statistic for detection of jumps in the mean is known to be sensitive to outliers. We thus propose a robust test based on the Wilcoxon two-sample test statistic. The asymptotic distribution of this test can be derived from a functional central limit theorem for two-sample U-statistics. We extend a theorem of Csörgő and Horváth to the case of dependent data.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Babbel, B.: Invariance principles for U-statistics and von Mises functionals. J. Stat. Plan. Inference 22, 337–354 (1989)
Billinsgley, P.: Convergence of Probability Measures. 2nd edn. Wiley, New York (1999)
Borovkova, S.A., Burton, R.M., Dehling, H.G.: Limit theorems for functionals of mixing processes with applications to U-statistics and dimension estimation. Trans. Am. Math. Soc. 353, 4261–4318 (2001)
Carlstein, E.: The use of subseries values for estimating the variance of a general statistic from a stationary sequence. Ann. Stat. 14, 1171–1179 (1986)
Csörgő, M., Horvath, L.: Invariance principles for changepoint problems. J. Multivar. Anal. 27, 151–168 (1988)
Csörgő, M., Horvath, L.: Limit Theorems in Change Point Analysis. Wiley, New York (1997)
Dehling, H.: Limit theorems for sums of weakly dependent Banach space valued random variables. Z. für Wahrscheinlichkeitstheorie und verwandte Gebiete 63, 393–432 (1983)
Dehling, H., Fried, R.: Asymptotic distribution of two-sample empirical U-quantiles with applications to robust tests for shifts in location. J. Multivar. Anal. 105, 124–140 (2012)
Dehling, H., Fried, R., Sharipov, O.Sh., Vogel, D., Wornowizki, M.: Estimation of the variance of partial sums of dependent processes. Stat. Probab. Lett. 83, 141–147 (2013)
Dehling, H., Rooch, A., Taqqu, M.S.: Nonparametric change-point tests for long-range dependent data. Scand. J. Stat. 40, 153–173 (2013)
Denker, M.: Asymptotic Distribution Theory in Nonparametrics Statistics. Vieweg Verlag, Braunschweig/Wiesbaden (1985)
Denker, M., Keller, G.: On U-statistics and v. Mises’ statistics for weakly dependent processes. Z. für Wahrscheinlichkeitstheorie und verwandte Gebiete 64, 505–522 (1983)
Denker, M., Keller, G.: Rigorous statistical procedures for data from dynamical systems. J. Stat. Phys. 44, 67–93 (1986)
Hoeffding, W.: A class of statistics with asymptotically normal distribution. Ann. Math. Stat. 19, 293–325 (1948)
Ibragimov, I.A., Linnik, Yu.V.: Independent and Stationary Sequences of Random Variables. Wolters-Noordhoff, Groningen (1971)
Leucht, A.: Degenerate U- and V-statistics under weak dependence: asymptotic theory and bootstrap consistency. Bernoulli 18, 552–585 (2012)
Rooch, A.: Change-point tests for long-range dependent data. Dissertation, Ruhr-Universität Bochum (2012)
Sen, P.K.: On the properties of U-statistics when the observations are not independent. I. Estimation of non-serial parameters in some stationary stochastic processes. Calcutta Stat. Assoc. Bull. 12, 69–92 (1963)
Sen, P.K.: Limiting behavior of regular functionals of empirical distributions for stationary ∗-mixing processes. Z. für Wahrscheinlichkeitstheorie und Verwandte Gebiete 25, 71–82 (1972)
Sharipov, O.Sh., Wendler, M.: Bootstrap for the sample mean and for U-statistics of mixing and near-epoch dependent processes. J. Nonparametric Stat. 24, 317–342 (2012)
Wooldridge, J.M., White, H.: Some invariance principles and central limit theorems for dependent heterogeneous processes. Econ. Theory 4, 210–230 (1988)
Yoshihara, K.-I.: Limiting behavior of U-statistics for stationary, absolutely regular processes. Z. für Wahrscheinlichkeitstheorie und verwandte Gebiete 35, 237–252 (1976)
Acknowledgements
The authors wish to thank the referees for their very careful reading of an earlier version of this manuscript, and for their many thoughtful comments that helped to improve the presentation of the paper. This research was supported by the Collaborative Research Center 823, Project C3 Analysis of Structural Change in Dynamic Processes, of the German Research Foundation DFG.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix: Some Auxiliary Results from the Literature
Appendix: Some Auxiliary Results from the Literature
In this section, we collect some known lemmas and theorems for weakly dependent data. We start with some results on the behaviour of partials sums:
Lemma 3 (Borovkova, Burton, Dehling [3], Lemma 2.23).
Let \((X_{k})_{k\in \mathbb{Z}}\) be a 1-approximating functional with constants (ak)k≥0 of an absolutely regular process with mixing coefficients (β(k))k≥0. Suppose moreover that EXi = 0 and that one of the following two conditions holds:
-
1.
X 0 is bounded a.s. and \(\sum _{k=0}^{\infty }(a_{k} +\beta (k)) <\infty.\)
-
2.
\(E\vert X_{0}\vert ^{2+\delta } <\infty\) and \(\sum _{k=0}^{\infty }(a_{k}^{ \frac{\delta }{ 1+\delta } } +\beta ^{ \frac{\delta }{1+\delta } }(k)) <\infty.\)
Then, as \(N \rightarrow \infty\) ,
and the sum on the r.h.s. converges absolutely.
Lemma 4 (Borovkova, Burton, Dehling [3], Lemma 2.24).
Let \((X_{k})_{k\in \mathbb{Z}}\) be a 1-approximating functional with constants (a k ) of an absolutely regular process with mixing coefficients (β(k)) k≥0 . Suppose moreover that EX i = 0 and that one of the following two conditions holds:
-
1.
X 0 is bounded a.s. and \(\sum _{k=0}^{\infty }k^{2}(a_{k} +\beta (k)) <\infty.\)
-
2.
\(E\vert X_{0}\vert ^{4+\delta } <\infty\) and \(\sum _{k=0}^{\infty }k^{2}(a_{k}^{ \frac{\delta }{ 3+\delta } } +\beta ^{ \frac{\delta }{4+\delta } }(k)) <\infty.\)
Then there exits a constant C such that
Theorem 4 (Borovkova, Burton, Dehling [3], Theorem 4).
Let \((X_{k})_{k\in \mathbb{Z}}\) be a 1-approximating functional with constants (a k ) k≥0 of an absolutely regular process with mixing coefficients (β(k)) k≥0 . Suppose moreover that EX i = 0, \(E\vert X_{0}\vert ^{4+\delta } <\infty\) and that
for some δ > 0. Then, as \(n \rightarrow \infty,\)
where \(\sigma ^{2} = EX_{0}^{2} + 2\sum _{j=1}^{\infty }E(X_{0}X_{j}).\) In case \(\sigma ^{2} = 0\) , \(\mathcal{N}(0,0)\) denotes the point mass at the origin. If X 0 is bounded, the CLT continues to hold if (77) is replaced by the condition that \(\sum _{k=0}^{\infty }k^{2}(a_{k} +\beta (k)) <\infty\) .
An important tool to derive asymptotic results for weakly dependent data are coupling methods. We will apply this method in the proof of Proposition 2.
Theorem 5 (Borovkova, Burton, Dehling [3], Theorem 3).
Let \((X_{n})_{n\in \mathbb{N}}\) be a 1-approximating functional with summable constants (a k ) k≥0 of an absolutely regular process with mixing rate (β(k)) k≥0 . Then given integers K,L and N, we can approximate the sequence of \((K + 2L,N)-\) blocks (B s ) s≥1 by a sequence of independent blocks \((B_{s}^{{\prime}})_{s\geq 1}\) with the same marginal distribution in such a way that
where \(\alpha _{L}:= \left (2\sum _{l=L}^{\infty }a_{l}\right )^{1/2}.\)
In statistical application, the question of how to estimate \(\sigma ^{2}\) is important. In the situation when the observations are a functional of α-mixing process, Dehling et al. [10] propose the estimation of the variance of partial sums of dependent processes by the subsampling estimator
with \(\hat{T}_{i}(l) =\sum _{ j=(i-1)l+1}^{il}F_{n}(X_{j})\) and \(\tilde{U}_{n} = \frac{1} {n}\sum _{j=1}^{n}F_{ n}(X_{j})\), where F n (⋅ ) is the empirical distribution function.
Theorem 6 (Dehling, Fried, Sharipov, Vogel, Wornowizki [9], Theorem 1.2).
Let (X k ) k≥1 be a stationary, 1-approximating functional of an α-mixing processes. Suppose that for some δ > 0, \(E\vert X_{1}\vert ^{2+\delta } <\infty\) , and that the mixing coefficients (α k ) k≥1 and the approximation constants (a k ) k≥1 satisfy
In addition, we assume that F is Lipschitz-continuous, that \(\alpha _{k} = O(n^{-8})\) and that \(a_{m} = O(m^{-12})\) . Then, as \(n \rightarrow \infty\) , \(l_{n} \rightarrow \infty\) and \(l_{n} = o(\sqrt{n})\) , we have \(\hat{D}_{n}\longrightarrow \sigma\) in L 2 .
To deal with the degenerate kernel g, we need to find upper bounds for the expectations \(E\left (g(X_{i_{1}},X_{j_{1}})g(X_{i_{2}},X_{j_{2}})\right )\), in terms of the maximal distance among the indices. Since 1 ≤ i 1 < i 2 ≤ [n λ] and \([n\lambda ] + 1 \leq j_{1} <j_{2} \leq n\), we get i 1 < i 2 < j 1 < j 2.
Lemma 5 (Dehling, Fried [8], Proposition 6.1).
Let (X n ) n≥1 be a 1-approximating functional with constants (a k ) k≥1 of an absolutely regular process with mixing coefficients (β(k)) k≥1 and let g(x,y) be a 1-continuous bounded degenerate kernel. Then we have
where \(S = \vert \sup _{x,y}g(x,y)\vert\) and \(k =\max \left \{i_{2} - i_{1},j_{1} - i_{2},j_{2} - j_{1}\right \}\) .
The following two results are useful for proving tightness of a stochastic process. The first one is used to control the fluctuation of maximum. Let ξ 1, …, ξ n be random variables, and define \(S_{k} =\xi _{1} +\ldots +\xi _{k}\) (S 0 = 0), and \(M_{n} =\max _{0\leq k\leq n}\vert S_{k}\vert\).
Theorem 7 (Billingsley [2], Theorem 10.2).
Suppose that β ≥ 0 and α > 1∕2 and that there exist nonnegative numbers u 1 ,…,u n such that for all positive λ
then for all positive λ
where K β,α is a constant depending only on β and α.
Theorem 8 (Billingsley [2], Theorem 8.4).
The sequence {Y n }, defined by
is tight if for each ε > 0 there exist a λ > 1 and a \(n_{0} \in \mathbb{N}\) such that for n ≥ n 0
Rights and permissions
Copyright information
© 2015 Springer Science+Business Media New York
About this chapter
Cite this chapter
Dehling, H., Fried, R., Garcia, I., Wendler, M. (2015). Change-Point Detection Under Dependence Based on Two-Sample U-Statistics. In: Dawson, D., Kulik, R., Ould Haye, M., Szyszkowicz, B., Zhao, Y. (eds) Asymptotic Laws and Methods in Stochastics. Fields Institute Communications, vol 76. Springer, New York, NY. https://doi.org/10.1007/978-1-4939-3076-0_12
Download citation
DOI: https://doi.org/10.1007/978-1-4939-3076-0_12
Publisher Name: Springer, New York, NY
Print ISBN: 978-1-4939-3075-3
Online ISBN: 978-1-4939-3076-0
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)