Electronic Journal of Statistics

Prediction of dynamical time series using kernel based regression and smooth splines

Raymundo Navarrete and Divakar Viswanath

Full-text: Open access


Prediction of dynamical time series with additive noise using support vector machines or kernel based regression is consistent for certain classes of discrete dynamical systems. Consistency implies that these methods are effective at computing the expected value of a point at a future time given the present coordinates. However, the present coordinates themselves are noisy, and therefore, these methods are not necessarily effective at removing noise. In this article, we consider denoising and prediction as separate problems for flows, as opposed to discrete time dynamical systems, and show that the use of smooth splines is more effective at removing noise. Combination of smooth splines and kernel based regression yields predictors that are more accurate on benchmarks typically by a factor of 2 or more. We prove that kernel based regression in combination with smooth splines converges to the exact predictor for time series extracted from any compact invariant set of any sufficiently smooth flow. As a consequence of convergence, one can find examples where the combination of kernel based regression with smooth splines is superior by even a factor of $100$. The predictors that we analyze and compute operate on delay coordinate data and not the full state vector, which is typically not observable.

Article information

Electron. J. Statist., Volume 12, Number 2 (2018), 2217-2237.

Received: December 2017
First available in Project Euclid: 23 July 2018

Permanent link to this document

Digital Object Identifier

Mathematical Reviews number (MathSciNet)

Zentralblatt MATH identifier

Creative Commons Attribution 4.0 International License.


Navarrete, Raymundo; Viswanath, Divakar. Prediction of dynamical time series using kernel based regression and smooth splines. Electron. J. Statist. 12 (2018), no. 2, 2217--2237. doi:10.1214/18-EJS1429. https://projecteuclid.org/euclid.ejs/1532333003

Export citation


  • [1] T.M. Adams and A.B. Nobel. On density estimation from ergodic processes., The Annals of Probability, 25:794–804, 1998.
  • [2] V. Araújo and I. Melbourne. Exponential decay of correlations for nonuniformly hyperbolic flows with a $c^1+\alpha $ stable foliation, including the classical Lorenz attractor., Ann. Henri Poincaré, 17 :2975–3004, 2015.
  • [3] A. Christmann and I. Steinwart. Consistency and robustness of kernel-based regression in convex risk minimization., Bernoulli, 3:799–819, 2007.
  • [4] C. de Boor., A Practical Guide to Splines. Springer, New York, revised edition, 2001.
  • [5] P.P.B. Eggermont and V.N. LaRiccia. Uniform error bounds for smoothing splines. In, High Dimensional Probability, volume 51 of IMS Lecture Notes-Monograph Series, pages 220–237. Institute of Mathematical Statistics, 2006.
  • [6] P.P.B. Eggermont and V.N. LaRiccia., Maximum Penalized Likelihood Estimation, volume II. Springer, New York, 2009. Springer Series in Statistics.
  • [7] I. Ekeland and R. Temam., Convex Analysis and Variational Problems. SIAM, Philadelphia, 1987.
  • [8] J. Doyne Farmer and J.J. Sidorowich. Predicting chaotic time series., Physical Review Letters, 59:845–848, 1987.
  • [9] L. Györfi, W. Härdle, P Sarda, and P. Vieu., Nonparametric Curve Estimation from Time Series, volume 60 of Lecture Notes in Statistics. Springer, New York, 1989.
  • [10] L. Györfi, M. Kohler, A. Krzyżak, and H. Walk., A Distribution-Free Theory of Nonparametric Regression. Springer, New York, 2002.
  • [11] H. Hang, Y. Feng, I. Steinwart, and J.A.K. Suykens. Learning theory estimates with observations from general stationary stochastic processes., Neural Computation, 28 :2853–2889, 2016.
  • [12] S. P. Lalley. Beneath the noise, chaos., The Annals of Statistics, 27:461–479, 1999.
  • [13] S. P. Lalley and A. B. Nobel. Denoising deterministic time series., Dynamics of PDE, 3:259–279, 2006.
  • [14] S.P. Lalley. Removing the noise from chaos plus noise. In A.I. Mees, editor, Nonlinear Dynamcs and Statistics, pages 233–244. Birkhäuser, Boston, 2001.
  • [15] D. Mattera and S. Haykin. Support vector machines for dynamic reconstrution of a chaotic system. In B. Schölkopf, C.J.C. Burges, and A.J. Smola, editors, Advances in Kernel Methods, pages 211–242. MIT Press, MA, 1999.
  • [16] K. McGoff, S. Mukherjee, and N. Pillai. Statistical inference for dynamical systems: A review., Statistics Surveys, 9:209–252, 2015.
  • [17] K. McGoff and A.B. Nobel. Empirical risk minimization and complexity of dynamical models., www.arxiv.org, 2016.
  • [18] S. Mukherjee, E. Osuna, and F. Girosi. Nonlinear prediction of chaotic time series using a support vector machine. In, Neural Networks for Signal Processing VII—Proceedings of the 1997 IEEE Workshop, pages 511–520. IEEE, 1997.
  • [19] K.-S. Müller, A.J. Smola, G. Rätsch, B. Schölkopf, J. Kohlmorgen, and V.N. Vapnik. Using support vector machines for time series prediction. In B. Scholkopf, C.J.C. Burges, and S. Mika, editors, Advances in Kernel Methods, pages 243–253. MIT Press, 1998.
  • [20] N.I. Sapankevych and R. Sankar. Time series prediction using support vector machines: a survey., IEEE Computational Intelligence Magazine, pages 24–38, May 2009.
  • [21] T. Sauer, J.A. Yorke, and M. Casdagli. Embedology., Journal of Statistical Physics, 65:579–616, 1991.
  • [22] B. Schölkopf and A.J. Smola., Learning with Kernels. MIT Press, MA, 2002.
  • [23] I. Steinwart. On the influence of the kernel on the consistency of support vector machines., Journal of Machine Learning Research, 2:67–93, 2001.
  • [24] I. Steinwart and M. Anghel. Consistency of support vector machines for forecasting the evolution of an unknown ergodic dynamical system from observations with unknown noise., The Annals of Statistics, 37:841–875, 2009.
  • [25] I. Steinwart, D. Hush, and C. Scovel. Learning from dependent observations., Journal of Multivariate Analysis, 100:175–194, 2009.
  • [26] C. Stone. Optimal rates of convergence for nonparametric regression., Annals of Statistics, 10 :1040–1063, 1982.
  • [27] K. Urbanowicz and J.A. Holyst. Noise-level estimation of time series using coarse-grained entropy., Physical Review E, 67 :046218, 2003.
  • [28] D. Viswanath, X. Liang, and K. Serkh. Metric entropy and the optimal prediction of chaotic signals., SIAM Journal on Applied Dynamical Systems, 12 :1085–1113, 2013.
  • [29] G. Wahba., Spline Models for Observational Data. SIAM, Philadelphia, 1990.
  • [30] D.-X. Zhou. Derivative reproducing properties for kernel methods in learning theory., Journal of Computational and Applied Mathematics, 220:456–463, 2008.