PROC COUNTREG Statement :: SAS/ETS(R) 14.1 User's Guide

Data Set Options

DATA=SAS-data-set

specifies the input SAS data set. If the DATA= option is not specified, PROC COUNTREG uses the most recently created SAS data set.

GROUPID=variable

specifies an identification variable when a panel data model is estimated. The identification variable is used as a cross-sectional ID variable.

NONORMALIZE

does not row-normalize the spatial weights matrix that is specified in the WMAT= option. By default, the spatial weights matrix is required to be row-normalized; that is, the spatial weights matrix has unit row sum. Equivalently, this means that $w(\mathbf{s}_{i},\mathbf{s}_{j})$ is normalized by multiplying it by $\frac{1}{\sum _{j=1}^ n w(\mathbf{s}_{i},\mathbf{s}_{j})}$ , where n is the total number of spatial units. If the NONORMALIZE option is specified, spatial weights are used "as is" except for $w(\mathbf{s}_{i},\mathbf{s}_{i})$ , which is always treated as 0. This implies that a spatial weight $w(\mathbf{s}_{i},\mathbf{s}_{j})$ cannot be missing for $i\neq j$ if the NONOMALIZE option is specified. If the NONOMALIZE option is not specified, missing spatial weights are replaced with 0.

WMAT=SAS-data-set

specifies the input SAS data set that contains spatial weights matrix. The spatial weights matrix is often known as the W matrix. The spatial weights $w(\mathbf{s}_{i},\mathbf{s}_{j})$ for two locations $\mathbf{s}_{i}$ and $\mathbf{s}_{j}$ must satisfy the following: $w(\mathbf{s}_{i},\mathbf{s}_{j}) \geq 0$ and $w(\mathbf{s}_{i},\mathbf{s}_{i})=0$ , where $i,j=1,2,\ldots ,n$ and n is the total number of spatial locations. However, it is not necessary that $w(\mathbf{s}_{i},\mathbf{s}_{j})=w(\mathbf{s}_{j},\mathbf{s}_{i})$ . In addition, any nonzero $w(\mathbf{s}_{i},\mathbf{s}_{i})$ is replaced with 0. For more information about missing spatial weights in W, see the section NONORMALIZE.

For a spatial weights data set that has n spatial units, the number of columns must be $n+1$ if the SPATIALID statement specifies a spatial ID variable for the purpose of matching observations. For more information, see the section SPATIALID Statement. However, if the SPATIALID statement is not specified, the number of rows and columns in the spatial weights data set must be equal.

Item Store Control Options

RESTORE=item-store-name: specifies the source item store for processing. An item-store-name consists of a one- or two-level name, as with SAS data sets. As with data sets, an item store is associated by default with the Work library, and any item stores that are created in this library are deleted when the SAS session concludes.

Output Data Set Options

OUTEST=SAS-data-set: writes the parameter estimates to the specified output data set.
COVOUT: writes the covariance matrix for the parameter estimates to the OUTEST= data set. This option is valid only if the OUTEST= option is specified.

Printing Options

CORRB: prints the correlation matrix of the parameter estimates. This option can also be specified in the MODEL statement.
COVB: prints the covariance matrix of the parameter estimates. This option can also be specified in the MODEL statement.
NOPRINT: suppresses all printed output.

Estimation Control Options

COVEST=value

specifies the type of covariance matrix of the parameter estimates. The quasi-maximum-likelihood estimates are computed using COVEST=QML. By default, COVEST=HESSIAN. The supported covariance types are as follows:

HESSIAN: specifies the covariance from the Hessian matrix.
OP: specifies the covariance from the outer product matrix.
QML: specifies the covariance from the outer product and Hessian matrices.

Plot Control Options

PLOTS<(global-plot-options)> < = plot-request | (plot-requests)>: requests that the COUNTREG procedure produce statistical graphics via the Output Delivery System, provided that ODS GRAPHICS has been enabled. For general information about ODS Graphics, see Chapter 21: Statistical Graphics Using ODS in SAS/STAT 14.1 User's Guide. The global-plot-options apply to all relevant plots that are generated by the COUNTREG procedure.

You can specify the following global-plot-options:

COUNTS(value1 <value2...>): supplies the plots PREDPROB and PREDPROFILE with particular values of the response variable. Each value should be a nonnegative integer. Nonintegers are rounded to the nearest integer. The value can also be a list of the form X TO Y BY Z. For example, COUNTS(0 1 2 TO 10 BY 2 15) specifies plotting for counts 0, 1, 2, 4, 6, 8, 10, and 15.
ONLY: suppresses the default plots. Only the plots that are specifically requested are produced.
UNPACKPANEL UNPACK: displays each graph separately. (By default, some graphs can appear together in a single panel.)

You can specify the following plot-requests:

ALL: requests that all plots appropriate for the particular analysis be produced.
AUTOCORR<(LAGS=n)>: displays the autocorrelation function plots of the parameters. This plot-request is available only for Bayesian analysis. The optional LAGS= suboption specifies the number (up to lag n) of autocorrelations to be plotted in the AUTOCORR plot. If this suboption is not specified, autocorrelations are plotted up to lag 50.
BAYESDIAG: displays the TRACE, AUTOCORR, and DENSITY plots. This plot-request is available only for Bayesian analysis.
BAYESSUM: displays the posterior distribution, prior distribution, and maximum likelihood estimates. This plot-request is available only for Bayesian analysis.
DENSITY<(FRINGE)>: displays the kernel density plots of the parameters. This plot-request is available only for Bayesian analysis. If you specify the FRINGE suboption, a fringe plot is created on the X axis of the kernel density plot.
DISPERSION: produces the overdispersion diagnostic plot.
NONE: suppresses all plots.
PREDPROB: produces the overall predictive probabilities of the specified count levels. You must also specify COUNTS in global-plot-options.
PREDPROFILE: produces the predictive probability profiles of specified count levels against model regressors. The regressor on the X axis is varied, whereas all other regressors are fixed at the mean of the observed data set.
PROFILELIKE: produces the profile likelihood functions of the model parameters. The model parameter on the X axis is varied, whereas all other parameters are fixed at their estimated maximum likelihood estimates.
TRACE<(SMOOTH)>: displays the trace plots of the parameters. This plot-request is available only for Bayesian analysis. The SMOOTH suboption displays a fitted penalized B-spline curve for each trace plot.
ZEROPROFILE | ZPPRO: produces the probability profiles of zero-inflation process selection and zero count prediction against model regressors. The regressor on the X axis is varied, whereas all other regressors are fixed at the mean of the observed data set.

Optimization Process Control Options

PROC COUNTREG uses the nonlinear optimization (NLO) subsystem to perform nonlinear optimization tasks. All the NLO options are available in the NLOPTIONS statement. For more information, see the NLOPTIONS Statement. In addition, you can specify the following option in the PROC COUNTREG statement:

METHOD=value

specifies the iterative minimization method to use. By default, METHOD=NRA.

CONGRA: specifies the conjugate-gradient method.
DBLDOG: specifies the double-dogleg method.
NMSIMP: specifies Nelder-Mead simplex method.
NRA: specifies the Newton-Raphson method.
NRRIDG: specifies the Newton-Raphson ridge method.
QN: specifies the quasi-Newton method.
TR: specifies the trust region method.

The COUNTREG Procedure