In addition to the output data sets, the VARIOGRAM procedure produces a variety of output objects. Most of these are produced depending on whether you specify either NOVARIOGRAM or LAGDISTANCE= and MAXLAGS= in the COMPUTE statement. The VARIOGRAM procedure output objects are the following:
a default “Number of Observations” table that displays the number of observations read from the input data set and the number of observations used in the analysis
a default map that shows the spatial distribution of the observations of the current variable in the VAR statement. The observations are displayed by default with circled markers whose color indicates the VAR value at the corresponding location.
a table with basic information about the lags and the extreme distance between data pairs, when NOVARIOGRAM is specified
a table that describes the distribution of data pairs in distance intervals, when NOVARIOGRAM is specified
a histogram plot of the pairwise distance distribution, when NOVARIOGRAM is specified). The plot also displays a reference line at a user-specified pairs frequency threshold when you specify the THRESHOLD= parameter in the PLOTS=PAIRS option. The option PLOTS=PAIRS(NOINSET) forces the informational inset that appears in the plot to hide.
empirical semivariogram details, when NOVARIOGRAM is not specified and LAGDISTANCE= and MAXLAGS= are specified. This table also includes the semivariance estimate variance and confidence limits when CL is specified, and estimates of the robust semivariance when ROBUST is specified.
plots of the appropriate empirical semivariograms, when NOVARIOGRAM is not specified and LAGDISTANCE= and MAXLAGS= are specified. If you perform the analysis in more than one direction simultaneously, the output is a panel that contains the empirical semivariogram plots for the specified angles. If the semivariograms are nonpaneled, then each plot includes in the lower part a needle plot of the contributing pairs distribution.
a table that provides autocorrelation statistics, when the options AUTOCORRELATION and LAGDISTANCE= are specified
the Moran scatter plot of the standardized observation values against the weighted averages of their neighbors, when the options PLOTS=MORAN, AUTOCORRELATION, and LAGDISTANCE= are specified
When you specify the MODEL statement and request a fit of a theoretical model to the empirical semivariogram, the VARIOGRAM procedure also produces the following default output:
a table with some general fitting information, in addition to the output item store if you have specified one with the STORE statement
a table with more specific information about the selected model’s parameters and their initial values
a table with general information about the optimization that provides the fitting parameters of the selected model
a table with the optimization process output and a table with the convergence status of the optimization process, if you have specified a single model to fit
a “Parameter Estimates” table with information about the fitted parameters estimates
a “Fit Summary” table that reports the fit quality of all models you requested to fit
plots of fitted theoretical semivariogram models. If you perform model fitting in more than one direction angle or for more than one variable in your DATA= data set, then the output is a panel that contains all fitted models for the respective directions or variables.
Additional output can be produced in model fitting if you specify a higher level of output detail with the DETAILS option in the MODEL statement. This output can be information tables for each separate model when you specify multiple models to fit, tables with more details about the optimization process, and the covariance and correlation matrices of the model parameter estimates. The complete listing of the PROC VARIOGRAM output follows in the section ODS Table Names and the section ODS Graph Names.