Dresden 2011 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

DY: Fachverband Dynamik und Statistische Physik

DY 38: Data Analysis and Stochastic Modeling I (jointly with UP)

DY 38.2: Talk

Thursday, March 17, 2011, 16:45–17:00, ZEU 255

An advanced method for the estimation of drift and diffusion coefficients from stochastic time series — •Christoph Honisch and Rudolf Friedrich — Institut für Theoretische Physik, Universität Münster, D-48149 Münster

We present a novel iterative method to estimate drift and diffusion coefficients from data of stationary univariate Markov processes X(t). These coefficients are defined as D⁽ⁿ⁾(x) = 1/n! lim_{τ → 0} 1/τ ⟨ (X(t+τ)−x)ⁿ|x=X(t) ⟩, where n=1,2 corresponds to drift and diffusion coefficients respectively. Our method overcomes the problem of performing the limit τ → 0 by taking advantage of a recently reported approach [1] to calculate exact finite sampling interval effects on the estimation of drift and diffusion. Therefore, good results are achieved in cases of sparsely sampled time series.

[1] St. J. Lade, Phys. Lett. A 373, 3705 (2009)