ESTIMATE <data>,<name of model>,L
estimates the parameters of a linear or a nonlinear regression model.
Parameter L is optional and gives the first line for the results in
the edit field.
The regression model has to be defined by a MODEL specification typed on
any two consecutive lines in the edit field as follows:
MODEL <name of model>
<regressand>=<function of regressors and parameters>
<data> may either be defined by the DATA specification in the edit
field or it may be a Survo data file.
The regressand and the regressors are notated by the names given as
labels for the columns in the DATA specification of the data set.
Unidentified words in the model are interpreted as parameters to be
estimated.
ESTIMATE is able to evaluate the expressions of first and second
derivatives of the model function with respect to parameters. This
information is employed in the model identification and in the numerical
computation of estimates.
The estimation process can be controlled by extra specifications typed
in the edit field.
Extra specifications and parameters in ESTIMATE:
1 = model specification (MODEL)
2 = estimation criterion (CRITERION)
3 = numerical (optimization) method (METHOD)
4 = initial values for parameters
5 = weight variable (WEIGHT)
6 = step length (STEP)
7 = residuals, predicted values
8 = selecting observations
9 = constants in the model function
A = displaying intermediate results
M = results in matrix files
D = more on data analysis
Help System (web edition)