Option set for bj
opt = bjOptions
opt = bjOptions(Name,Value)
creates
the default options set for opt
= bjOptionsbj
.
creates
an option set with the options specified by one or more opt
= bjOptions(Name,Value
)Name,Value
pair
arguments.
Specify optional commaseparated pairs of Name,Value
arguments.
Name
is the argument
name and Value
is the corresponding
value. Name
must appear
inside single quotes (' '
).
You can specify several name and value pair
arguments in any order as Name1,Value1,...,NameN,ValueN
.
'InitialCondition'
— Handling of initial conditions'auto'
(default)  'zero'
 'estimate'
 'backcast'
Handling of initial conditions during estimation, specified as one of the following strings:
'zero'
— The initial conditions
are set to zero.
'estimate'
— The initial
conditions are treated as independent estimation parameters.
'backcast'
— The initial
conditions are estimated using the best least squares fit.
'auto'
— The software chooses
the method to handle initial conditions based on the estimation data.
'Focus'
— Estimation focus'prediction'
(default)  'simulation'
 'stability'
 vector  matrix  linear systemEstimation focus that defines how the errors e between the measured and the modeled outputs are weighed at specific frequencies during the minimization of the prediction error, specified as one of the following:
'prediction'
— Automatically
calculates the weighting function as a product of the input spectrum
and the inverse of the noise spectrum. The weighting function minimizes
the onestepahead prediction. This approach typically favors fitting
small time intervals (higher frequency range). From a statisticalvariance
point of view, this weighting function is optimal. However, this method
neglects the approximation aspects (bias) of the fit.
This option focuses on producing a good predictor and does not
enforce model stability. Use 'stability'
when you
want to ensure a stable model.
'simulation'
— Estimates
the model using the frequency weighting of the transfer function that
is given by the input spectrum. Typically, this method favors the
frequency range where the input spectrum has the most power. This
method provides a stable model.
'stability'
— Same as 'prediction'
,
but with model stability enforced.
Passbands — Row vector or matrix containing frequency values that define desired passbands. For example:
[wl,wh] [w1l,w1h;w2l,w2h;w3l,w3h;...]
where wl
and wh
represent
lower and upper limits of a passband. For a matrix with several rows
defining frequency passbands, the algorithm uses union of frequency
ranges to define the estimation passband.
Passbands are expressed in rad/TimeUnit
,
where TimeUnit
is the time units of the estimation
data.
SISO filter — Specify a SISO linear filter in one of the following ways:
A singleinputsingleoutput (SISO) linear system
{A,B,C,D}
format, which specifies
the statespace matrices of the filter
{numerator, denominator}
format,
which specifies the numerator and denominator of the filter transfer
function
This option calculates the weighting function as a product of
the filter and the input spectrum to estimate the transfer function.
To obtain a good model fit for a specific frequency range, you must
choose the filter with a passband in this range. The estimation result
is the same if you first prefilter the data using idfilt
.
'EstCovar'
— Control whether to generate parameter covariance datatrue
(default)  false
Controls whether parameter covariance data is generated, specified
as true
or false
.
If EstCovar
is true
,
then use getcov
to fetch the
covariance matrix from the estimated model.
'Display'
— Specify whether to display the estimation progress'off'
(default)  'on'
Specify whether to display the estimation progress, specified as one of the following strings:
'on'
— Information on model
structure and estimation results are displayed in a progressviewer
window.
'off'
— No progress or results
information is displayed.
'InputOffset'
— Removal of offset from timedomain input data during estimation[]
(default)  vector of positive integers  matrixRemoval of offset from timedomain input data during estimation,
specified as the commaseparated pair consisting of 'InputOffset'
and
one of the following:
A column vector of positive integers of length Nu, where Nu is the number of inputs.
[]
— Indicates no offset.
NubyNe matrix
— For multiexperiment data, specify InputOffset
as
an NubyNe matrix. Nu is
the number of inputs, and Ne is the number of experiments.
Each entry specified by InputOffset
is
subtracted from the corresponding input data.
'OutputOffset'
— Removal of offset from timedomain output data during estimation[]
(default)  vector  matrixRemoval of offset from timedomain output data during estimation,
specified as the commaseparated pair consisting of 'OutputOffset'
and
one of the following:
A column vector of length Ny, where Ny is the number of outputs.
[]
— Indicates no offset.
NybyNe matrix
— For multiexperiment data, specify OutputOffset
as
a NybyNe matrix. Ny is
the number of outputs, and Ne is the number of
experiments.
Each entry specified by OutputOffset
is
subtracted from the corresponding output data.
'Regularization'
— Options for regularized estimation of model parametersstructureOptions for regularized estimation of model parameters. For more information on regularization, see Regularized Estimates of Model Parameters.
Regularization
is a structure with the following
fields:
Lambda
— Constant that determines
the bias versus variance tradeoff.
Specify a positive scalar to add the regularization term to the estimation cost.
The default value of zero implies no regularization.
Default: 0
R
— Weighting matrix.
Specify a vector of nonnegative numbers or a square positive semidefinite matrix. The length must be equal to the number of free parameters of the model.
For blackbox models, using the default value is recommended.
For structured and greybox models, you can also specify a vector
of np
positive numbers such that each entry denotes
the confidence in the value of the associated parameter.
The default value of 1 implies a value of eye(npfree)
,
where npfree
is the number of free parameters.
Default: 1
Nominal
— The nominal value
towards which the free parameters are pulled during estimation.
The default value of zero implies that the parameter values
are pulled towards zero. If you are refining a model, you can set
the value to 'model'
to pull the parameters towards
the parameter values of the initial model. The initial parameter values
must be finite for this setting to work.
Default: 0
'SearchMethod'
— Search method used for iterative parameter estimation'auto'
(default)  'gn'
 'gna'
 'lm'
 'lsqnonlin'
 'grad'
Search method used for iterative parameter estimation, specified as one of the following strings:
'gn'
— The subspace GaussNewton
direction. Singular values of the Jacobian matrix less than GnPinvConst*eps*max(size(J))*norm(J)
are
discarded when computing the search direction. J is
the Jacobian matrix. The Hessian matrix is approximated by J^{T}J.
If there is no improvement in this direction, the function tries the
gradient direction.
'gna'
— An adaptive version
of subspace GaussNewton approach, suggested by Wills and Ninness [1]. Eigenvalues less than gamma*max(sv)
of
the Hessian are ignored, where sv are the singular
values of the Hessian. The GaussNewton direction is computed in the
remaining subspace. gamma has the initial value InitGnaTol
(see Advanced
for
more information). This value is increased by the factor LMStep
each
time the search fails to find a lower value of the criterion in less
than 5 bisections. This value is decreased by the factor 2*LMStep
each
time a search is successful without any bisections.
'lm'
— Uses the LevenbergMarquardt
method so that the next parameter value is pinv(H+d*I)*grad
from
the previous one. H is the Hessian, I is
the identity matrix, and grad is the gradient. d is
a number that is increased until a lower value of the criterion is
found.
'lsqnonlin'
— Uses lsqnonlin
optimizer from Optimization Toolbox™ software.
You must have Optimization Toolbox installed to use this option.
This search method can handle only the Trace criterion.
'grad'
— The steepest descent
gradient search method.
'auto'
— The algorithm chooses
one of the preceding options. The descent direction is calculated
using 'gn'
, 'gna'
, 'lm'
,
and 'grad'
successively at each iteration. The
iterations continue until a sufficient reduction in error is achieved.
'SearchOption'
— Option set for the search algorithmsearch option setOption set for the search algorithm with fields that depend
on the value of SearchMethod
.
SearchOption structure when SearchMethod is specified as 'gn', 'gna', 'lm', 'grad', or 'auto'
Field Name  Description  

Tolerance  Minimum percentage difference (divided by 100) between
the current value of the loss function and its expected improvement
after the next iteration. When the percentage of expected improvement
is less than Default:  
MaxIter  Maximum number of iterations during lossfunction minimization.
The iterations stop when Setting Use Default:  
Advanced  Advanced search settings. Specified as a structure with the following fields:

SearchOption structure when SearchMethod is specified as ‘lsqnonlin'
Field Name  Description 

TolFun  Termination tolerance on the loss function that the software minimizes to determine the estimated parameter values. The
value of Default: 
TolX  Termination tolerance on the estimated parameter values. The
value of Default: 
MaxIter  Maximum number of iterations during lossfunction minimization.
The iterations stop when The
value of Default: 
Advanced  Options set for For more information, see the Optimization Options table in Optimization Options. Use 
'Advanced'
— Additional advanced optionsstructureAdditional advanced options, specified as a structure with the following fields:
ErrorThreshold
— Specifies
when to adjust the weight of large errors from quadratic to linear.
Errors larger than ErrorThreshold
times the
estimated standard deviation have a linear weight in the criteria.
The standard deviation is estimated robustly as the median of the
absolute deviations from the median and divided by 0.7
.
For more information on robust norm choices, see section 15.2 of [2].
ErrorThreshold = 0
disables
robustification and leads to a purely quadratic criterion. When estimating
with frequencydomain data, the software sets ErrorThreshold
to
zero. For timedomain data that contains outliers, try setting ErrorThreshold
to 1.6
.
Default: 0
MaxSize
— Specifies the
maximum number of elements in a segment when inputoutput data is
split into segments.
MaxSize
must be a positive integer.
Default: 250000
StabilityThreshold
— Specifies
thresholds for stability tests.
StabilityThreshold
is a structure with the
following fields:
s
— Specifies the location
of the rightmost pole to test the stability of continuoustime models.
A model is considered stable when its rightmost pole is to the left
of s
.
Default: 0
z
— Specifies the maximum
distance of all poles from the origin to test stability of discretetime
models. A model is considered stable if all poles are within the distance z
from
the origin.
Default: 1+sqrt(eps)
AutoInitThreshold
— Specifies
when to automatically estimate the initial condition.
The initial condition is estimated when
$$\frac{\Vert {y}_{p,z}{y}_{meas}\Vert}{\Vert {y}_{p,e}{y}_{meas}\Vert}>\text{AutoInitThreshold}$$
y_{meas} is the measured output.
y_{p,z} is the predicted output of a model estimated using zero initial states.
y_{p,e} is the predicted output of a model estimated using estimated initial states.
Applicable when InitialCondition
is 'auto'
.
Default: 1.05
opt
— Options set for bj
bjOptions
option setOption set for bj
, returned
as an bjOptions
option set.
opt = bjOptions;
Create an options set for bj
using zero initial conditions for estimation. Set Display
to 'on'
.
opt = bjOptions('InitialCondition','zero','Display','on');
Alternatively, use dot notation to set the values of opt
.
opt = bjOptions; opt.InitialCondition = 'zero'; opt.Display = 'on';
[1] Wills, Adrian, B. Ninness, and S. Gibson. "On GradientBased Search for Multivariable System Estimates". Proceedings of the 16th IFAC World Congress, Prague, Czech Republic, July 3–8, 2005. Oxford, UK: Elsevier Ltd., 2005.
[2] Ljung, L. System Identification: Theory for the User. Upper Saddle River, NJ: PrenticeHall PTR, 1999.