System identification

System identification is a field that uses statistical methods to build mathematical models of dynamical systems from measured data, describing how a system behaves in the time or frequency domain. It also covers the design of experiments meant to generate data that is informative for fitting such models, and supports model reduction, with a major application in control systems, where it underlies modern data driven control. The field uses three broad kinds of models: white box models built from first principles, such as Newton's equations, which are often impractically complex; black box models, the most common approach, which relate inputs to outputs without examining internal mechanics; and grey box models, a hybrid that combines partial knowledge of the system with experimental data to estimate the remaining unknown parameters. Methods can also be classed as input-output, which use both input and output data and are generally more accurate, or output-only, which rely on output measurements alone. This description is adapted from Wikipedia contributors under CC BY-SA 4.0; changes were made. https://creativecommons.org/licenses/by-sa/4.0/