PRINT Subcommand (PROXMAP command) 

[/PRINT [DATA [([RAW]
              [COMBINED]
              [DERIVED])]]
        [WEIGHTS [([RAW]
                 [OBJECTS]
                 [DERIVED])]]
        [INITIAL [([COORDINATES]
                   [VARDIRECTIONS]
                   [PROPERTYDIRECTIONS]
                   [STRESS]
                   [DIAGNOSTICS])]]
        [MODEL [([COORDINATES]   
                 [DIRECTIONS]
                 [SPACEWEIGHTS])]]
        [ATTRIBUTES [([RAW]
                    [TRANSFORMED]
                    [FITTED]
                    [ATTRIBUTE_WEIGHTS]
                    [DIRECTIONS])]]
        [PROPERTIES [([RAW]
                      [TRANSFORMED]
                      [FITTED]
                      [DIRECTIONS])]]
        [CRITERIA [([SUMMARY** ]
                    [DIAGNOSTICS]
                    [ITERATIONS]
                    [STRESS])]]]

PRINT controls which tables are displayed in the output. The summary of the analysis, CRITERIA(SUMMARY), is always displayed by default, but all other output is optional.

  • All main keywords (for example, CONFIGURATION, DISTANCES) can be followed by one or more sub keywords in parentheses to fine-tune the output.
  • When the main keyword is specified, at least one sub keyword should be given.
  • Inapplicable keywords are silently ignored. For instance, /PRINT DATA(DERIVED) has no effect if the input data are proximities (as specified on /DATA = PROXIMITIES).

Labeling in output:

  • Object labels in tables and plots are governed by the LABELS keyword of the /OBJECTID subcommand.
  • If a variable name is specified with LABELS, objects are labeled with its values or value labels (if available).

The PRINT subcommand includes the following main keywords and sub keywords.

Keyword Possible values and description
DATA Describes the input data and derived proximities.

RAW displays the original proximities. For multivariate data, it displays the original variables per source.

DERIVED displays the proximities derived from multivariate variables.

COMBINED displays the proximities combined over sources (in the multiple sources case).
WEIGHTS

RAW displays the original proximity weights.

OBJECTS indicates the case weights for proximity data.

DERIVED indicates the derived proximity weights.
INITIAL Information about the initial configuration.

COORDINATES indicates the coordinates of the initial configuration (from classical scaling or INITIAL VARIABLES).

VARDIRECTIONS: For multivariate data, coordinates of variables such as, vectors for non-nominal, points for nominal measurement level.

PROPERTYDIRECTIONS indicates the coordinates of property variables such as vectors for non-nominal and points for nominal measurement level.

STRESS indicates the full decomposition of stress over sources, dimensions, and objects.

DIAGNOSTICS indicates the initial diagnostics based on the INITIAL subcommand specification.

For Classical: Eigenvalues, eigenvectors, and their statistics.

For Variables: Optimal scaling factor and initial normalized stress.
MODEL Final (common space) configuration.

COORDINATES: Coordinates of objects in the common space; includes individual spaces if multiple sources.

DIRECTIONS: Coordinates of multivariate variables in the common space. It also shows variance accounted for per variable and per dimension.

SPACEWEIGHTS: Weights applied to spaces (sources or dimensions), depending on the model type specified in MODEL.
ATTRIBUTES Applicable when variables are specified on ATTRIBUTES.

RAW: Original attribute variables.

TRANSFORMED: Transformed attribute variables.

FITTED: Fitted attribute variables.

ATTRIBUTE_WEIGHTS: Weights for attribute variables in restricting the common space.

DIRECTIONS: Coordinates of attribute variable vectors in the common space; variance accounted for is also displayed.
PROPERTIES

Applicable when variables are specified on PROPERTIES.

RAW: Original property variables.

TRANSFORMED: Transformed property variables.

DIRECTIONS: Coordinates of property variable vectors in the common space and/or points for nominal variables. Variance accounted for is also displayed.
CRITERIA

Diagnostic and algorithmic statistics.

SUMMARY: Elapsed time, number of iterations, final normalized and Kruskal-1 stress, and error messages (if any).

DIAGNOSTICS: Additional statistics including Kruskal’s stress formula 1, Alienation coefficient, Dispersion accounted for, Tucker’s congruence coefficient, and Pearson and Spearman correlations.

ITERATIONS: Summarized iteration history: normalized stress, stress difference, and coordinate change per iteration.

STRESS: Full decomposition of final stress (also appears under STRESS above).

Example

GET FILE ’States.sav’.
PROXMAP
 /SOURCEID DATA = income life NAME = $SRC1
 /SOURCEID DATA = school fail NAME = $SRC2
 /SOURCEID DATA = popul freeze NAME = $SRC3
 /DATA = MULTIVARIATE
 /MODEL WEIGHTED
 /PRINT DATA(COMBINED) CRITERIA(ITERATIONS) MODEL(COORDINATES SPACEWEIGHTS)

Here, a weighted Euclidean model is specified with three sources.

The optional table output consists of a table of combined proximities across sources, the summarized history of iterations, the coordinates in all spaces (both general (common) and model-specific (individual) spaces) and the model space weights (which are dimension weights with this model).