User`s guide

4 Nonlinear Black-Box Model Identification
You can specify different nonlinearity estimators for different output channels
by setting
InputNonlinearity or Outpu tNon linearity to an object array.
For example, the follow ing co de estimates a tw o-input H ammerstein-Wiener
model, where
sigmoidnet and pwl ine ar are the two input nonlinearities
and there is no output nonlinearity:
m = nlhw(data,[nb,nf,nk],...
[sigmoidnet;pwlinear],...
[])
Alternatively, you can construct the model first and then estimate models
parameters using the following commands:
m0 = idnlhw([nb,nf,nk],[sigmoidnet;pwlinear],[]);
m = nlhw(data,m0);
For detailed information about nlhw and idnlhw, see the corresponding
reference pages.
For more information about validating your models, see Chapter 8, “Model
Analysis”.
You can also use
pem to re fine parameter estimates of an existing
Hammerstein-Wiener model, as described in “Refining Nonlinear Black-Box
Models” on page 4-28.
Improving Estimation Results U sing Initial States
If your estimated Hammerstein-Wiener model provides a poor fittomeasured
data, you can estimate the model again using initial states estimated from
the data. By default, the initial states corresponding to the linear block of the
Hammerstein-Wiener model are ze ro.
To specify estimating initial states during model estimation, you can use
the following syntax:
m0 = idnlhw([nb,nf,nk],[sigmoidnet;pwlinear],[]);
m = nlhw(data,m0,'InitialState','e');
4-22