Skip to main content
Log in

Wavelet-Type Expansion of the Generalized Rosenblatt Process and Its Rate of Convergence

  • Published:
Journal of Fourier Analysis and Applications Aims and scope Submit manuscript

Abstract

Pipiras introduced in the early 2000s an almost surely and uniformly convergent (on compact intervals) wavelet-type expansion of the classical Rosenblatt process. Yet, the issue of estimating, almost surely, its uniform rate of convergence remained an open question. The main goal of our present article is to provide an answer to it in the more general framework of the generalized Rosenblatt process, under the assumption that the underlying wavelet basis belongs to the class due to Meyer. The main ingredient of our strategy consists in expressing in a non-classical (new) way the approximation errors related with the approximation spaces of a multiresolution analysis of \(L^2({{\mathbb {R}}}^2)\). Such a non-classical expression may also be of interest in its own right.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

References

  1. Abry, P., Pipiras, V.: Wavelet-based synthesis of the Rosenblatt process. Signal Process. 86, 2326–2339 (2006)

    Article  Google Scholar 

  2. Arras, B.: A white noise approach to stochastic integration with respect to the Rosenblatt process. Potential Anal. 43(4), 547–591 (2015)

    Article  MathSciNet  Google Scholar 

  3. Ayache, A.: Lower bound for local oscillations of Hermite processes. Stoch. Process. Appl., to appear

  4. Ayache, A., Taqqu, M.S.: Rate optimality of wavelet series approximations of fractional Brownian motion. J. Fourier Anal. Appl. 9(5), 451–471 (2003)

    Article  MathSciNet  Google Scholar 

  5. Bai, S., Taqqu, M.S.: Generalized Hermite processes, discrete chaos and limit theorems. Stoch. Process. Appl. 124(4), 1710–1739 (2014)

    Article  MathSciNet  Google Scholar 

  6. Bai, S., Taqqu, M.S.: Structure of the third moment of the generalized Rosenblatt distribution. Stat. Probab. Lett. 94, 144–152 (2014)

    Article  MathSciNet  Google Scholar 

  7. Bai, S., Taqqu, M.S.: Convergence of long-memory discrete k-th order Volterra processes. Stoch. Process. Appl. 125(5), 2026–2053 (2015)

    Article  MathSciNet  Google Scholar 

  8. Bai, S., Taqqu, M.S.: Behavior of the generalized Rosenblatt process at extreme critical exponent values. Ann. Probab. 25(2), 1278–1324 (2017)

    Article  MathSciNet  Google Scholar 

  9. Bai, S., Taqqu, M.S.: The impact of the diagonals of polynomial forms on limit theorems with long memory. Bernoulli 23(1), 710–742 (2017)

    Article  MathSciNet  Google Scholar 

  10. Bardet, J.-M., Tudor, C.A.: A wavelet analysis of the Rosenblatt process: chaos expansion and estimation of the self-similarity parameter. Stoch. Process. Appl. 120(12), 2331–2362 (2010)

    Article  MathSciNet  Google Scholar 

  11. Bojdecki, T., Gorostiza, L.G., Talarczyk, A.: From intersection local time to the Rosenblatt process. J. Theoret. Probab. 28, 1227–1249 (2015)

    Article  MathSciNet  Google Scholar 

  12. Daubechies, I.: Ten Lectures on Wavelets. Society for Industrial and Applied Mathematics, Philadelphia (1992)

    Book  Google Scholar 

  13. Dobrushin, R.L., Major, P.: Non-central limit theorems for non-linear functional of Gaussian fields. Zeitschrift für Wahrscheinlichkeitstheorie und Verwandte Gebiete 50(1), 22–52 (1979)

    Article  Google Scholar 

  14. Embrechts, P., Maejima, M.: Self-similar Processes. Academic Press, New York (2003)

    MATH  Google Scholar 

  15. Janson, S.: Gaussian Hilbert Spaces. Cambridge Tracts in Mathematics. Cambridge University Press, Cambridge (1997)

    Book  Google Scholar 

  16. Maejima, M., Tudor, C.A.: Selfsimilar processes with stationary increments in the second Wiener chaos. Probab. Math. Stat. 32(1), 167–186 (2012)

    MathSciNet  MATH  Google Scholar 

  17. Maejima, M., Tudor, C.A.: On the distribution of the Rosenblatt process. Stat. Probab. Lett. 83(6), 1490–1495 (2013)

    Article  MathSciNet  Google Scholar 

  18. Meyer, Y.: Wavelets and Operators, vol. 37. Cambridge University Press, Cambridge (1992)

    MATH  Google Scholar 

  19. Nualart, D.: The Malliavin Calculus and Related Topics. Springer, Berlin (2006). Number 2

    MATH  Google Scholar 

  20. Peccati, G., Taqqu, M.S.: Moments, Cumulants and Diagrams. Springer, New York (2011)

    Book  Google Scholar 

  21. Pipiras, V.: Wavelet-type expansion of the Rosenblatt process. J. Fourier Anal. Appl. 10(6), 599–634 (2004)

    Article  MathSciNet  Google Scholar 

  22. Pipiras, V., Taqqu, M.S.: Long-Range Dependence and Self-similarity. Cambridge University Press, Cambridge (2017)

    Book  Google Scholar 

  23. Rosenblatt, M.: Independence and dependence. In: Proc. 4th Berkeley Symp. Math. Stat. Probab., vol. 2, pp. 431–443. University California Press, Berkeley (1961)

  24. Samko, S.G., Kilbas, A.A., Marichev, O.I.: Fractional Integrals and Derivatives: Theory and Applications. Gordon and Breach Science Publishers, New York (1993)

    MATH  Google Scholar 

  25. Samorodnitsky, G., Taqqu, M.S.: Stable Non-Gaussian Random Variables: Stochastic Models with Infinite Variance. Chapman and Hall, New York (1994)

    MATH  Google Scholar 

  26. Taqqu, M.S.: Weak convergence to fractional Brownian motion and to the Rosenblatt process. Probab. Theory Relat. Fields 31(4), 287–302 (1975)

    MathSciNet  MATH  Google Scholar 

  27. Taqqu, M.S.: Convergence of integrated processes of arbitrary Hermite rank. Zeitschrift für Wahrscheinlichkeitstheorie und Verwandte Gebiete 50(1), 53–83 (1979)

    Article  MathSciNet  Google Scholar 

  28. Taqqu, M.S.: The Rosenblatt process. In: Davis, R.A., Lii, K., Politis, D.N. (eds.) Selected Works of Murray Rosenblatt, Selected Works in Probability and Statistics, pp. 29–45. Springer, New York (2011)

    Google Scholar 

  29. Veillette, M.S., Taqqu, M.S.: Properties and numerical evaluation of the Rosenblatt distribution. Bernoulli 19(3), 982–1005 (2013)

    Article  MathSciNet  Google Scholar 

Download references

Acknowledgements

The authors are grateful to the anonymous reviewers for their valuable comments and remarks which have led to improvements of the manuscript. This work has been partially supported by the Labex CEMPI (ANR-11-LABX-0007-01) and the GDR 3475 (Analyse Multifractale).

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Antoine Ayache.

Additional information

Communicated by Christian Houdre.

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix

Appendix

Lemma 4.1

For every \((x,y)\in {{\mathbb {R}}}_+^2\), one has

$$\begin{aligned} \log (3+x+y)\le \log (3+x)\log (3+y). \end{aligned}$$
(4.1)

Moreover, for any fixed positive real number T there exists a constant \(c>0\) such that, for all \(x\in {\mathbb {R}}_+\), the following inequality holds:

$$\begin{aligned} \log (3+x+2^xT)\le c (1+x). \end{aligned}$$
(4.2)

The proof of Lemma 4.1 is standard and easy this is why it has been omitted.

Lemma 4.2

For any fixed real number \(L> 1\), there exists a constant \(c>0\) such that, for all \(j\in {{\mathbb {Z}}}\) and for each \(s\in {{\mathbb {R}}}\), one has:

$$\begin{aligned} \sum _{k\in {{\mathbb {Z}}}}\frac{\sqrt{\log (3+|j|+|k|)}}{(3+|2^js-k|)^L}\le c\,\sqrt{\log \big (3+|j|+ 2^j |s| \big )}. \end{aligned}$$
(4.3)

Proof of Lemma  4.2

Setting \(m=k-\lfloor 2^js\rfloor \), where \(\lfloor 2^js \rfloor \) denotes the integer part of \(2^j s\), and using the triangle inequality and (4.1), one obtains that

$$\begin{aligned} \sum _{k\in {{\mathbb {Z}}}}\frac{\sqrt{\log (3+|j|+|k|)}}{(3+|2^js-k|)^L}&=\sum _{m\in {{\mathbb {Z}}}}\frac{\sqrt{\log \big (3+|j|+|m+\lfloor 2^js\rfloor |)}}{(3+|2^js-\lfloor 2^js\rfloor -m|)^L}\\&\le \sqrt{\log \big (3+|j|+ 2^j |s| \big )}\sum _{m\in {{\mathbb {Z}}}}\frac{\sqrt{\log (4+|m|)}}{(3+|2^js-\lfloor 2^js\rfloor -m|)^L}. \end{aligned}$$

Then, noticing that

$$\begin{aligned} 3+|2^js-\lfloor 2^js\rfloor -m|\ge 2+|m|, \end{aligned}$$

one gets that

$$\begin{aligned} \sum _{k\in {{\mathbb {Z}}}}\frac{\sqrt{\log (3+|j|+|k|)}}{(3+|2^js-k|)^L}\le c\,\sqrt{\log (3+|j|+2^j |s| \big )}, \end{aligned}$$

where the constant

$$\begin{aligned} c:=\sum _{m\in {{\mathbb {Z}}}}\frac{\sqrt{\log (4+|m|)}}{(2+|m|)^L}<+\infty . \end{aligned}$$

\(\square \)

Lemma 4.3

For each fixed real number \(L> 1\), there exists a constant \(c>0\) such that, for every \(t\in {{\mathbb {R}}}_+\), for all \(s\in [0,t]\) and for any \(j\in {{\mathbb {Z}}}_+\), one has

$$\begin{aligned} \sum _{k\in D_j^3(t)}\frac{\sqrt{\log (3+j+|k|)}}{(3+|2^js-k|)^L}\le c (j+1)2^{-j (L-1)(1-a)}\sqrt{\log (3+t)}\,, \end{aligned}$$
(4.4)

where \(D_j^3(t)\) is defined through (2.14) and (2.11).

Proof of Lemma 4.3

In view of (2.14) and (2.11), one has \( D_j^3(t)=D_j^{3,+}(t)\cup D_j^{3,-}(t), \) where \(D_j^{3,+}(t)\) and \(D_j^{3,-}(t)\) are the two disjoint sets defined as:

$$\begin{aligned} D_j^{3,+}(t)=\left\{ k\in {{\mathbb {Z}}}, \ k>2^jt+2^{j(1-a)}\right\} \end{aligned}$$

and

$$\begin{aligned} D_j^{3,-}(t)=\left\{ k\in {{\mathbb {Z}}}, \ k<-2^{j(1-a)}\right\} . \end{aligned}$$

Thus, one gets that

$$\begin{aligned} \sum _{k\in D_j^3(t)}\frac{\sqrt{\log (3+j+|k|)}}{(3+|2^js-k|)^L}= \sum _{k>2^jt+2^{j(1-a)}}\frac{\sqrt{\log (3+j+k)}}{(3+|2^js-k|)^L} +\sum _{k<-2^{j(1-a)}}\frac{\sqrt{\log (3+j+|k|)}}{(3+|2^js-k|)^L}.\nonumber \\ \end{aligned}$$
(4.5)

Let us now provide an appropriate upper bound for the first term in the right-hand side of (4.5). One denotes by \(\lfloor \cdot \rfloor \) the integer part function. Using the change of variable \(m=k-\lfloor 2^jt\rfloor \), the triangle inequality, (4.1), the inequality \(\lfloor 2^jt\rfloor -2^js >-1\), the inequality \(\log (3+m)\le 2+m\), and the fact that \(x\mapsto (1+x)^{-L}\sqrt{\log (2+x)}\) is a decreasing function on \({{\mathbb {R}}}_+\), one obtains that

$$\begin{aligned} \sum _{k>2^jt+2^{j(1-a)}}\frac{\sqrt{\log (3+j+k)}}{(3+|2^js-k|)^L}&=\sum _{k>2^jt+2^{j(1-a)}}\frac{\sqrt{\log (3+j+k)}}{(3+k-2^js)^L} \nonumber \\&=\sum _{m>2^jt-\lfloor 2^jt\rfloor +2^{j(1-a)}}\frac{\sqrt{\log (3+j+\lfloor 2^jt\rfloor +|m|)}}{(3+\lfloor 2^jt\rfloor -2^js+m)^L} \nonumber \\&\le \sqrt{\log (3+j+2^j t)}\sum _{m>2^{j(1-a)}}\frac{\sqrt{\log (3+m)}}{(2+m)^L}\nonumber \\&\le \sqrt{\log (3+j+2^j t)}\int _{2^{j(1-a)}}^{+\infty }\frac{\sqrt{\log (2+x)}}{(1+x)^{L}}\,dx\nonumber \\&\le c_1\sqrt{(j+1)\log (3+t)}\,\,2^{-j(L-1)(1-a)}\sqrt{\log \big (2+2^{j(1-a)}\big )}\nonumber \\&\le c_2 (j+1) 2^{-j(L-1)(1-a)}\sqrt{\log (3+t)}\, , \end{aligned}$$
(4.6)

where \(c_1\) and \(c_2\) are two positive finite constants not depending on j, t, s and a. Similarly to (4.6), it can be shown that

$$\begin{aligned} \sum _{k<-2^{j(1-a)}}\frac{\sqrt{\log (3+j+|k|)}}{(3+|2^js-k|)^L}\le c_3 2^{-j(L-1)(1-a)}\sqrt{(j+1)\log (3+j)}\,, \end{aligned}$$
(4.7)

where \(c_3\) is a positive finite constant not depending on j, t, s and a. Finally, putting together (4.5), (4.6), (4.7) and (4.2), it follows that (4.4) holds. \(\square \)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Ayache, A., Esmili, Y. Wavelet-Type Expansion of the Generalized Rosenblatt Process and Its Rate of Convergence. J Fourier Anal Appl 26, 51 (2020). https://doi.org/10.1007/s00041-020-09757-3

Download citation

  • Received:

  • Revised:

  • Published:

  • DOI: https://doi.org/10.1007/s00041-020-09757-3

Keywords

Mathematics Subject Classification

Navigation