Location: Jabara Hall, Room 372

Data-Driven Discovery of Dynamics with Uncertainty Quantification Prof. Chris Wikle, University of Missouri

Abstract:

Environmental processes are governed by complex dynamical interactions across multiple scales and multiple processes. The specification of realistic models for such dynamical processes can be facilitated by mechanistically motivated models, with effective regularization. However, often the time scales or multi-process complexity of such processes preclude the effective specification of a specific mechanism by which to motivate the model. Recently, there has been interest to use data-driven discovery methods to learn the fundamental dynamical mechanisms present in data. These approaches have shown promise in simulations of low-dimensional systems without a great deal of noise or missingness in the observations. Here, we present a fully Bayesian statistical approach for dynamic discovery that accommodates uncertain observations that can be applied even with relatively short time series. The method is illustrated on simulated systems and applied to real-world data examples.

Joint work with Josh North (PhD Student) and Erin Schliep (Assoc. Professor)

Please join us for refreshments at 2:30 p.m. in Jabara Hall, Room 353.