The NLIN Procedure

ODS Graphics

Subsections:

Convergence Status Table

Statistical procedures use ODS Graphics to create graphs as part of their output. ODS Graphics is described in detail in Chapter 21: Statistical Graphics Using ODS.

Before you create graphs, ODS Graphics must be enabled (for example, by specifying the ODS GRAPHICS ON statement). For more information about enabling and disabling ODS Graphics, see the section Enabling and Disabling ODS Graphics in Chapter 21: Statistical Graphics Using ODS.

The overall appearance of graphs is controlled by ODS styles. Styles and other aspects of using ODS Graphics are discussed in the section A Primer on ODS Statistical Graphics in Chapter 21: Statistical Graphics Using ODS.

PROC NLIN assigns a name to each graph it creates using ODS. You can use these names to refer to the graphs when using ODS. The graphs that are controlled by the PLOTS option in the PROC NLIN statement are listed in Table 81.10, those that are controlled by the options in PROFILE statement are in Table 81.11.

Table 81.10: Graphs Controlled by the PLOTS option in the PROC NLIN Statement

ODS Graph Name	Plot Description	PLOTS Option
ContourFitPlot	Contour fit plot for models with two regressors	FIT
FitPlot	Fit plot for models with one regressor	FIT
FitDiagnosticsPanel	Panel of fit diagnostics	DIAGNOSTICS
LeveragePlot	Tangential and Jacobian leverages versus observation number	DIAGNOSTICS
LocalInfluencePlot	Local influence versus observation number	DIAGNOSTICS
ObservedByPredictedPlot	Dependent variable versus predicted values	DIAGNOSTICS(UNPACK)
ProjectedResidualHistogram	A histogram of the projected residuals	DIAGNOSTICS(UNPACK)
RawResidualExpectationPlot	Raw residual expectation versus predicted values	DIAGNOSTICS(UNPACK)
RawResidualHistogram	A histogram of the raw residuals	DIAGNOSTICS(UNPACK)
ResidualBoxPlot	A box plot of the raw and projected residuals	DIAGNOSTICS(UNPACK)
ResidualPanel	A panel of the raw and projected residuals versus the regressors	RESIDUALS
ResidualPlot	A plot of the raw and projected residuals versus the regressors	RESIDUALS(UNPACK)
ResidualByPredictedPlot	Raw and projected residuals versus the predicted values	DIAGNOSTICS(UNPACK)
RStudentByJacLeveragePlot	Standardized raw and projected residuals versus Jacobian leverage	DIAGNOSTICS(UNPACK)
RStudentByPredictedPlot	Standardized raw and projected residuals versus the predicted values	DIAGNOSTICS(UNPACK)
RStudentByTanLeveragePlot	Standardized raw and projected residuals versus tangential leverage	DIAGNOSTICS(UNPACK)

Table 81.11: Graphs Controlled by the PROFILE Statement

ODS Graph Name	Plot Description	PROFILE Option
ConfidenceCurve	Parameter value versus t value	CONFCURV
JackknifePlot	Absolute relative percentage difference versus observation number	JACKKNIFE
ProfiletPlot	Likelihood ratio pivotal statistic versus Wald pivotal statistic	TPLOT

Table 81.12: Graphs Controlled by the BOOTSTRAP Statement

ODS Graph Name	Plot Description	BOOTSTRAP Option
BootstrapHistPlot	A histogram of bootstrap parameter estimates	BOOTPLOTS(HIST)
BootstrapScatterPlot	Pairwise scatter plot of bootstrap parameter estimates	BOOTPLOTS(SCATTER)

Convergence Status Table

The "Convergence Status" table can be used to programmatically check the status of an estimation. This table contains the Status variable that takes on the value 0, 1, 2, or 3. If Status takes on a value less than 3, the convergence criterion was met. Specifically, the values mean the following:

Status=0: indicates that the convergence criterion was met and no warning or error messages were issued during the PROC NLIN run. Also, no notes that could indicate a problem with the model were issued.
Status=1: indicates that the convergence criterion was met and notes were written to the log that might indicate a problem with the model.
Status=2: indicates that the convergence criterion was met and one or more warning messages were produced during the PROC NLIN run.
Status=3: indicates that the convergence criterion was not met.

The following sample program demonstrates how the "Convergence Status" table can be used:

ods output ConvergenceStatus=ConvStatus;
proc nlin data=YourData;
   parameters a=1 b=1 c=1;
   model wgt = a + x / (b*y+c*z);
run;

data _null_;
   set ConvStatus;
   if status > 0 then put "A problem occurred";
run;