Anticipation of critical transitions in complex systems with applications to real world systems
Basic data for this project
Type of project: Own resources project
Duration at the University of Münster: since 01/10/2019
Description
In the course of the project, Bayesian statistics and modern numerical methods are used to estimate stability measures on data windows in order to characterise their temporal evolution. Special emphasis is placed on the modelling and estimation of a stochastic differential equation, where the deterministic part can provide information about the resilience and the diffusive part could yield information about the probability of noise-induced destabilisation. The development of the algorithm is accompanied by tests on simulated data and, if available, transferred to empirical data. The algorithm will also be optimised in terms of computing time and tailored and extended to possible characteristics of the time series under investigation.
Keywords: early warning signals; leading indicators; nonlinear physics; time series analysis; Bayesian statistics
Website of the project: https://github.com/MartinHessler/antiCPy
Project management at the University of Münster
| Thiele, Uwe | Professur für Theoretische Physik (Prof. Thiele) Center for Nonlinear Science Center for Multiscale Theory and Computation |
Research associates from the University of Münster
| Heßler, Martin | Center for Nonlinear Science |
| Kamps, Oliver | Center for Nonlinear Science |