Sparse Identification of Nonlinear Dynamics

Sparse Identification of Nonlinear Dynamics, known as SINDy, is a data-driven method for discovering the governing equations of a dynamical system directly from measurements, developed by Steven L. Brunton, J. Nathan Kutz and Joshua L. Proctor and published in 2016 in the Proceedings of the National Academy of Sciences. The method takes time-series snapshots of a system's behavior and performs a sparsity-promoting regression, such as LASSO, against a large library of candidate nonlinear functions, on the assumption that most physical systems are governed by only a small number of dominant terms once expressed in a suitable coordinate system. The result is an interpretable mathematical model of the system's dynamics, and the technique has been applied in areas including fluid dynamics and the study of complex biological networks. This description is adapted from Wikipedia contributors under CC BY-SA 4.0; changes were made. https://creativecommons.org/licenses/by-sa/4.0/