INITIAL Subcommand (PROXMAP command)
[/INITIAL {CLASSICAL** [{CONSTANT = {‘0.0’* }}]}]
{value }
{$SYSMIS}
{RANKINGS }
{VARIABLES = varlist }
The INITIAL
subcommand specifies the starting configuration for the main
iterative algorithm. If not explicitly defined, the default is classical scaling, also known as
Principal Coordinates Analysis (PCoA).
Several options are available for the initial configuration:
CLASSICAL
-
This is the default. It performs classical scaling on the input or derived proximities.
CONSTANT: When
DATA=PROXIMITIES
, you may also specify a value (usually negative) for an additive constant, or request an estimate of it by the keyword ESTIMATE. Both the value and ESTIMATE need to be in quotation marks. The fixed constant must not result in negative proximities.RANKINGS: With this keyword, classical scaling is applied to the rank orders of the proximities (the smallest proximity receives rank 1, the next smallest receives rank 2, and so on). No additive constant is used. This approach reduces the influence of outliers by focusing on relative ordering rather than raw values.
VARIABLES
-
User-provided coordinates can be specified as the initial configuration. The specified variables must be available in the active dataset and contain no missing values. The number of variables must be at least equal to the number of dimensions (
/CRITERIA=DIMENSIONS(p)
) used in the analysis. If more variables are provided than dimensions, the coordinates are rotated to principal axes, and the first p principal components are retained.
Example
GET FILE ’Ekman.sav’
PROXMAP /SOURCEID DATA = nm434 nm445 nm465 nm472 nm490 nm504 nm537 nm555 nm584 nm600 nm610 nm628 nm651
nm674
/DATA = PROXIMITY
/INITIAL variables = Var_dim1 Var_dim2.
This example uses variables Var_dim1 and Var_dim2 as coordinates in the initial configuration.