Statistical Inference for Stochastic Processes

, Volume 15, Issue 3, pp 193–223 | Cite as

Non-parametric estimation of the diffusion coefficient from noisy data

  • Emeline Schmisser


We consider a diffusion process (X t ) t ≥ 0, with drift b(x) and diffusion coefficient σ(x). At discrete times t k  = k δ for k from 1 to M, we observe noisy data of the sample path, \({Y_{k\delta}=X_{k\delta}+\varepsilon_{k}}\) . The random variables \({\left(\varepsilon_{k}\right)}\) are i.i.d, centred and independent of (X t ). The process (X t ) t ≥ 0 is assumed to be strictly stationary, β-mixing and ergodic. In order to reduce the noise effect, we split data into groups of equal size p and build empirical means. The group size p is chosen such that Δ = p δ is small whereas M δ is large. Then, the diffusion coefficient σ 2 is estimated in a compact set A in a non-parametric way by a penalized least squares approach and the risk of the resulting adaptive estimator is bounded. We provide several examples of diffusions satisfying our assumptions and we carry out various simulations. Our simulation results illustrate the theoretical properties of our estimators.


Diffusion coefficient Model selection Noisy data Non-parametric estimation Stationary distribution 

Mathematics Subject Classification

Primary 62G08 Secondary 62M05 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Barlow MT, Yor M (1982) Semimartingale inequalities via the Garsia-Rodemich-Rumsey lemma, and applications to local times. J Funct Anal 49(2): 198–229MathSciNetCrossRefzbMATHGoogle Scholar
  2. Barron A, Birgé L, Massart P (1999) Risk bounds for model selection via penalization. Probab Theory Relat Fields 113(3): 301–413CrossRefzbMATHGoogle Scholar
  3. Comte F, Rozenholc Y (2002) Adaptive estimation of mean and volatility functions in (auto-)regressive models. Stoch Process Appl 97(1): 111–145MathSciNetCrossRefzbMATHGoogle Scholar
  4. Comte F, Rozenholc Y (2004) A new algorithm for fixed design regression and denoising. Ann Inst Stat Math 56(3): 449–473MathSciNetCrossRefzbMATHGoogle Scholar
  5. Comte F, GenonCatalot V, Rozenholc Y (2007) Penalized nonparametric mean square estimation of the coefficients of diffusion processes. Bernoulli 13(2): 514–543MathSciNetCrossRefzbMATHGoogle Scholar
  6. Gloter A (2000) Discrete sampling of an integrated diffusion process and parameter estimation of the diffusion coefficient. ESAIM Probab Stat 4: 205–227MathSciNetCrossRefzbMATHGoogle Scholar
  7. Gloter A, Jacod J (2001) Diffusions with measurement errors. II. Optimal estimators. ESAIM Probab Stat 5:243–260 (electronic)Google Scholar
  8. Hall P, Heyde CC (1980) Martingale limit theory and its application. Academic Press, New YorkGoogle Scholar
  9. Jacod J, Li Y, Mykland PA, Podolskij M, Vetter M (2009) Microstructure noise in the continuous case: the pre-averaging approach. Stoch Process Appl 119:2249–2276Google Scholar
  10. Meyer Y (1990) Ondelettes et opérateurs. I Actualités Mathématiques [Current mathematical topics]. Hermann, Paris (Wavelets)Google Scholar
  11. Pardoux E, Veretennikov AY (2001) On the Poisson equation and diffusion approximation I. Ann Probab 29(3): 1061–1085MathSciNetCrossRefzbMATHGoogle Scholar
  12. Rosenbaum M (2007) etude de quelques problemes d’estimation statistique en finance. Ph.D. thesis, Universite Paris-Est et CRESTGoogle Scholar
  13. Schmisser E (2011) Nonparametric drift estimation from noisy data. Stat Decis 28(2): 119–150MathSciNetCrossRefzbMATHGoogle Scholar
  14. Zhang L, Mykland PA, Aït-Sahalia Y (2005) A tale of two time scales: determining integrated volatility with noisy high-frequency data. J Am Stat Assoc 100(472): 1394–1411CrossRefzbMATHGoogle Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2012

Authors and Affiliations

  1. 1.Laboratoire Paul PainlevéUniversité Lille 1Villeneuve d’AscqFrance

Personalised recommendations