27.03.2013 Views

SPSS® 12.0 Command Syntax Reference

SPSS® 12.0 Command Syntax Reference

SPSS® 12.0 Command Syntax Reference

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

1048 NLR<br />

• The two COMPUTE statements on the DERIVATIVES transformations list calculate the<br />

derivatives for the parameters A and B. If either one had been omitted, NLR would have<br />

calculated it numerically.<br />

• NLR specifies STOP as the dependent variable. It is not necessary to specify SPEED as the<br />

independent variable since it has been used in the model and derivatives programs.<br />

MODEL PROGRAM <strong>Command</strong><br />

The MODEL PROGRAM command assigns initial values to the parameters and signifies the<br />

beginning of the model program. The model program specifies the nonlinear equation<br />

chosen to model the data. There is no default model.<br />

• The model program is required and must precede the [C]NLR command.<br />

• The MODEL PROGRAM command must specify all parameters in the model program.<br />

Each parameter must be individually named. Keyword TO is not allowed.<br />

• Parameters can be assigned any acceptable SPSS variable name. However, if you intend<br />

to write the final parameter estimates to a file with the OUTFILE subcommand, do not use<br />

the name SSE or NCASES (see the OUTFILE subcommand on p. 1051).<br />

• Each parameter in the model program must have an assigned value. The value can be<br />

specified on MODEL PROGRAM or read from an existing parameter data file named on the<br />

FILE subcommand.<br />

• Zero should be avoided as an initial value because it provides no information on the scale<br />

of the parameters. This is especially true for CNLR.<br />

• The model program must include at least one command that uses the parameters and the<br />

independent variables (or preceding transformations of these) to calculate the predicted<br />

value of the dependent variable. This predicted value defines the nonlinear model. There<br />

is no default model.<br />

• By default, the program assumes that PRED is the name assigned to the variable for the<br />

predicted values. If you use a different variable name in the model program, you must<br />

supply the name on the PRED subcommand (see the PRED subcommand on p. 1052).<br />

• In the model program, you can assign a label to the variable holding predicted values and<br />

also change its print and write formats, but you should not specify missing values for this<br />

variable.<br />

• You can use any computational commands (such as COMPUTE, IF, DO IF, LOOP, END<br />

LOOP, END IF, RECODE, or COUNT) or output commands (WRITE, PRINT, or XSAVE) in<br />

the model program, but you cannot use input commands (such as DATA LIST, GET, MATCH<br />

FILES, or ADD FILES).<br />

• Transformations in the model program are used only by [C]NLR, and they do not affect the<br />

working data file. The parameters created by the model program do not become a part of<br />

the working data file. Permanent transformations should be specified before the model<br />

program.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!