System Identification
Anders Hansson, Martin Andersen · 2023
System identification is about learning models for dynamical systems. We define a regression problem for learning/estimating the dynamical system, and specifically we define it as a maximum likelihood problem. We discuss in detail how the parameters can be estimated by solving a nonlinear least squares problem. Recurrent neural networks and temporal convolutional neural networks are shown to be generalizations of the linear dynamical models to nonlinear dynamical models. The chapter is finished off with a discussion on experiment design for system identification.