User`s guide

Identifying Input-Output Polynomial Models
Estimating Model Orders at the Command Line
You can estimate model orders using the struc, arxstruc,andsels truc
commands in combination.
If you are working with a multiple-output system, you m u s t use
struc,
arxstruc,andselstruc comm ands for each output. In this case, you must
subreference the correct output channel in your estima t ion and validation
data sets.
For each estimation, you use two independent data sets—an estimation data
set and a validation data set. These independent data set can be from different
experiments, or you can select these data sets from a single experiment. For
more information about subreferencing data, see “Subreferencing iddata
Objects” on page 1-55 and “Subreferencing idfrd Objects” on page 1-70.
For an example of estimating model orders for a multiple-input system, see
“Estimatin g Dela ys in the Multiple-Input System” in System Identification
Toolbox Getting Started G uide.
struc. The
struc command creates a matrix of possible model-order
combinations for a specified range o f n
a
, n
b
,andn
k
values.
For example, the follow ing command defines the range of model orders and
delays
na=2:5, nb=1:5,andnk=1:5:
NN = struc(2:5,1:5,1:5))
Note struc applies only to single-input/single-output models. If you
have multiple inputs and want to use
struc, apply this command to one
input-output pair at a time.
arxstruc. The arxstruc command takes the output from stru c,estimates
an ARX model for each model o rder, and compares the model output to the
measured output.
arxstruc returns the loss for each model, which is the
normalized sum of squared prediction errors.
3-53