Model Structure Selection
In system identification, there are some different classes of the model that are distinguishable by their complexities. The examples are parametric model versus non-parametric model and discrete-time model versus continuous-time model. Some models are defined by given forms or equations, and these models are called parametric model. These models are dependent on some parameters and these parameters are unknown and need to be estimated. Nonparametric models are usually data measurement based and are referred to situations and experiments in which the outcome needs to be categorised. Such models are sometimes described by curves, and these curves carry some information about the characteristic properties of the system.
Choosing a model structure is usually the first step towards its estimation. There are various possibilities for structure – state-space, transfer functions and polynomial forms such as ARX, ARMAX, OE, BJ etc.