Parametric Models

The Parametric Models dialog is opened by chosing the corresponding item in the pop-up menu Estimate in the ident window.

The dialog box gives access to all parametric modeling in the toolbox and operates on the Working Data set in the main ident window.

Usage:First, choose a model structure using the pop-up menu. Then enter the orders of the model in the Orders: edit box, and hit the Return key. Possibly change the default Model name. Finally press the Estimate button to perform the estimation and enter the resulting model into the Model board.

The model structures supported by the dialog include:

ARX
The simplest model structure that is often a first choice. Enter in Order field: na nb nk, the orders and input delays.
ARMAX, OE, BJ
Covers a wide family of noise models for single output systems. Enter in Order field numerator and denominator orders. If you are uncertain about the interpretation, use the Order Editor
State Space
General linear state-space models. Enter in Order field the state space order. Optionally you may also enter the input delays within brackets. With the N4SID choice the order may be a range (like 1:10) which allows you to choose the order from a plot.
By Initial Model
Uses an arbitrary model structure created by previous estimation or by special Toolbox commands.
Process Models
Simple time-continuous process models are handled by another dialog. Choose 'Process Models...' in the Estimation pop-up menu.

Focus, Disturbance Model, Initial State and Covariance.

Iteration Information.

Order selection and the Order Edition.

Help topics.


More on the Parametric Models Dialog

More on the Parametric Models Dialog

Focus
Offers the choices Prediction, Estimation, Filter, and Stability.
'Prediction'
gives a standard prediction error estimation method. This has optimal statistical properties but from an approximation point,it typically favors a model fit at high frequencies.
'Simulation'
approximates the dynamics of the model (the transfer function from measured inputs to outputs) in a norm that is given by the input spectrum. For the noise model, a prediction error method is used, while the dynamics model is kept constant.
'Filter'
gives an additional weighting for the dynamics model fit, by the frequency contents of a user defined filter. When this option is chosen, a filter dialog window opens and you can design standard Butterworth filters that are used for the model fit.
'Stability'
forces the identified model to be stable.
Disturbance Model
If a model is written as a transfer function y = G u + H e, the rational funtion H is the disturbance model.

For ARX, ARMAX, OE and BJ models, this field indicates whether a disturbance model is estimated or fixed to 1 ('No disturbance model.').

For State-space models there is an option between estimating a noise model, which corresponds to the (Kalman gain) matrix K, or to fix it to zero, which gives H = 1.

Initial State
The models require certain initial signal values to compute model outputs, the initial state. For systems with slowly decaying responses, it may be very important to handle these initial states in a good way. The popupmenu gives you the following options:
Auto
This is the default. An automatic choice between the alternatives below is made, guided by the data properties.
Zero
All necessary initial values are taken as zero.
Estimate
The unknown intial values are treated as parameters, which are estimated.
Backcast
The initial values are estimated directly with a backwards filtering method due to T. Knudsen.
Covariance
Covariance = 'Estimate' is the default and normal choice. The the uncertainty of the estimated model will also be computed, and the confidence region of any Model View will also be diplayed whan asked for in the Options menu. For large models, the handling of uncertainty meausures may dominate the computation time.
Covariance = 'None' suppresses the estimation and handling of uncertainty measures. In particular for high order N4SID models this may be a useful option. Order Selection and Order Editor

Order selection and Order editor

Order editor...
Opens a dialog box that allows easy definition of the chosen model structure. Also provides detailed help about the various model structures.
Order selection
The order selection feature is available for state-space models, estimated using N4SID, and for single-output ARX models using the ARX option. Pushing this button fills out the Order field with default order ranges. Then push Estimate, and a special plot will open where models of different orders can be selected by clicking. Iteration Information

Iteration Information

Iteration control...
When applicable, opens a dialog box which provides information about the progress of iterative estimation algorithms. It also provides access to several options which govern the iterative process.

NOTE 1:
The ARX and State Space options allow many models to be generated simultaneously. Simply define the orders to be vectors using the : operator.
NOTE 2:
The Orders: edit box can accept a variable name from the MATLAB workspace. The variable must either contain valid orders or be an IDMODEL object.
NOTE 3:
The available options and orders are automatically adjusted to the number of inputs and outputs in the selected working data set.

Help topics.

(file idparest.htm)