Regensburg 2019 – scientific programme

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

DY: Fachverband Dynamik und Statistische Physik

DY 63: Modeling and Data Analysis

DY 63.6: Talk

Friday, April 5, 2019, 11:15–11:30, H20

Estimating model parameters by attractor comparison — •Ulrich Parlitz — Max Planck Institute for Dynamics and Self-Organization, Göttingen, Germany — Institute for Nonlinear Dynamics, Georg-August-Universität, Göttingen, Germany

Time series based methods for estimating parameters of dynamical models are discussed which are based on a comparison of the asymptotic dynamics given by (reconstructed) attractors underlying the observed data and those generated by the model. The similarity of both distributions of (reconstructed) states is quantified by different measures of (dis-)similarity or "distance" (e.g., ordinal pattern distributions, nearest neighbor distances) and a good set of model parameters minimizes their "distance" or dissimilarity. We present different implementations of this approach for parameter estimation and evaluate their performance with several dynamical examples.