Estimating Models

The pop-up menu Estimate provides various methods for estimating models based on the Working Data set. The models will be inserted into the Models board.

The pop-up has the following choices:

Parametric models...
Opens a dialog, which allows you to generate dynamic linear models with different structures, orders, and delays. A model structure characterizes the relationship between input and output data and between unknown noise sources and the output data. Supported model structures include ARX, ARMAX, Output Error (OE), Box-Jenkins (BJ), State-Space, and others.
Process models...
Opens a dialog, which allows you to simple continuous time dynamic linear models characterized by Static gain, time constants and time delays.
Spectral model...
Estimates the system's frequency response from the data using either Fourier Transform techniques or the Blackman-Tukey approach.
Correlation model...
Estimates the system's transient response (impulse response) by correlating filtered versions of the input-output data.
Quick start
Performs correlation analysis and spectral analysis. Computes a default ARX model with a delay heuristically determined based on the estimated impulse reponse of the system as well as a default order state-space model.

Help topics.

(file idestm.htm)