These functions are used by the Nonlinear Regression dialog.
menuNls
calls
tabSummary.nls and
tabPredict.nls if
summary and prediction results are requested.
a formula which specifies the nonlinear regression model.
nlsobj
an object that inherits from class
nls.
OPTIONAL ARGUMENTS:
data
a data frame in which to do the computations. In addition to the usual data
variables, the data frame may contain parameters (set, typically, by using the
assignment form of
parameters or
param)
that establish initial values for the model parameters.
start
a list or numerical vector.
Although it is optional, use of
start is recommended for unambiguous
specification of the parameters.
If
start is omitted,
the assumption is that any names occurring in
formula
that are not variables in the data frame are parameters.
The list form of
start allows the individual parameter names
to refer to subsets of the parameters of arbitrary length.
If a numeric starting vector is supplied, the named parameters must
each be of length 1.
In the case of partially linear models
(
plinear.p = T), only the nonlinear parameters should be supplied.
trace
if
TRUE, details of the iterations are printed.
maxiter
the maximum number of iterations during fitting.
tolerance
the tolerance for the convergence criterion in the algorithm. This is
a relative offset criterion that measures the numerical imprecision in
the parameter estimates compared to the statistical variability.
Smaller values of this will require more iterations while larger values
will result in convergence being declared earlier.
minscale
the minimum factor by which to shrink the default step size in an attempt
to decrease the sum of squares.
plinear.p
if
TRUE, the Golub-Pereyra algorithm
for partially linear least-squares models is used.
print.short.p
if
TRUE, a short summary of the nonlinear model is printed.
This output is from the function
print.nls.
print.long.p
if
TRUE, a long summary of the nonlinear model is printed.
This output is from the function
summary.nls.
save.name
a character string for the name of the data frame to save the
fit and residuals in.
If data frame with this name already exists in database 1 and it has the
appropriate number of rows then the saved values will be appended
to the data frame.
If the object already exist in database 1 and it is not a data frame
or it does not have the appropriate number of rows then a new name
is created by appending a number to
save.name and the results are
saved in the data frame with the new name.
save.fit.p
if
TRUE, the fitted values from the regression are saved in the
data frame
save.name.
save.resid.p
f
TRUE, the working residuals are saved in the data frame
save.name.
The working residuals are the response minus the fitted value.
newdata
a data frame to use for computing predictions.
It must contain the same names as the terms in the right side of the
formula for the model.
If missing, the predictions for the original data are computed.
predobj.name
a character string for the name of the data frame to save the
predictions, standard errors and confidence intervals in.
If data frame with this name already exists in database 1 and it has the
appropriate number of rows then the values will be appended
to the data frame.
If the object already exist in database 1 and it is not a data frame
or it does not have the appropriate number of rows then a new name
is created by appending a number to
predobj.name and the values are
saved in data frame with the new name.
predict.p
if
TRUE, the predicted values are saved in
predobj.name.
ci.p
if
TRUE, lower and upper confidence limits will be stored in the
predobj.name The column names will be "xx % L.C.L." and "xx% U.C.L"
where xx is the value specified in
conf.level.
These confidence limits are for the mean response and are computed as
the prediction plus/minus t-value * standard error.
se.p
if
TRUE, the pointwise standard errors for the predictions will be
stored in
predobj.name.
conf.level
the confidence level to use when computing confidence intervals.
VALUE:
invisibly returns an object of class
nls.
See the
nls.object help file for details.
SIDE EFFECTS:
Printed output will be displayed if requested.
Plots will be drawn if requested.
The objects
save.name and
predobj.name will be created or appended
to if fitted values, residuals or predictions are saved.