ARX model estimation using instrumental variable method with arbitrary instruments
sys = ivx(data,[na
nb nk],x)
sys = ivx(data,[na
nb nk],x,max_size)
estimates an
ARX polynomial model, sys
= ivx(data
,[na
nb nk]
,x
)sys
, using the instrumental
variable method with arbitrary instruments. The model is estimated
for the time series data data
. [na
nb nk]
specifies the ARX structure orders of the A and B polynomials
and the input to output delay, expressed in the number of samples.
An ARX model is represented as:
$$A(q)y(t)=B(q)u(tnk)+v(t)$$
specifies
the maximum size of matrices formed during estimation.sys
= ivx(data
,[na
nb nk]
,x
,max_size
)

Estimation data. The data can be:
When using frequencydomain data, the number of outputs must be 1. 

ARX model orders. For more details on the ARX model structure, see 

Instrument variable matrix.
The instruments used are analogous to the regression vector,
with 

Maximum matrix size.
Specify Default: 250000 

ARX model that fits the estimation data, returned as a discretetime Information about the estimation results and options used is
stored in the
For more information on using 
Use iv4
first
for IV estimation to identify ARX polynomial models where the instruments x
are
chosen automatically. Use ivx
for nonstandard
situations. For example, when there is feedback present in the data,
or, when other instruments need to be tried. You can also use iv
to
automatically generate instruments from certain custom defined filters.
[1] Ljung, L. System Identification: Theory for the User, page 222, Upper Saddle River, NJ, PrenticeHall PTR, 1999.