Figure 1

Methodology for sparse identification of nonlinear dynamics (SINDy) from data. First, data is generated from a dynamical system, in this case a biological network. The time series of data is synthesized into a nonlinear function library, and the terms in this library are related to the time derivative by an overdetermined linear regression problem. Enforcing sparsity ensures that only a small number of coefficients are nonzero, identifying the few active terms in the dynamics that are needed to model the system.