1: \begin{abstract}

2: We study efficiency of non-parametric estimation of diffusions (stochastic differential equations driven by Brownian motion) from long stationary trajectories.

3: First, we introduce estimators based on conditional expectation which is motivated by the definition of drift and diffusion coefficients. 

4: These estimators involve time- and space-discretization parameters for computing expected values from discretely-sampled stationary data.

5: Next, we analyze consistency and mean squared error of these estimators depending on 

6: computational parameters.

7: We derive relationships between the number of observational points, time- and space-discretization parameters

8: in order to achieve the optimal speed of convergence and minimize computational complexity. 

9: We illustrate our approach with numerical simulations. 

10: 

11: \noindent

12: Keywords: {Stochastic differential equations, non-parametric estimation, conditional expectation}

13: %\PACS{05.10.Gg \and 02.50.Tt \and 05.40.Jc}

14: \end{abstract}

15: