DATA Subcommand (PROXMAP command)


 [/DATA {MULTIVARIATE**} ] [= {EUCLIDEAN     }]
                             {STANDARDIZED**}
       {PROXIMITIES    }

The DATA subcommand specifies the type of data the cases and variables in the data file represent. The choice is between MULTIVARIATE (default) and PROXIMITIES.

MULTIVARIATE
The data is multivariate and represent measures of objects on variables. With this keyword, the variable list on the SOURCEID subcommands specify the variables from which to derive a proximity matrix. Distances between objects are computed to create this matrix, considering the measurement level of the variables (see the Data input subsection in the Overview section for details), and the distance function.

The distance function can be chosen following the MULTIVARIATE keyword with keywords:

EUCLIDEAN in which Euclidean distances are computed between cases.

STANDARDIZED in which Standardized Euclidean distances are computed between cases, which means computing Euclidean distances by using standardized variables. Standardization equalizes the contribution of each variable (Standardization to a sum of squares of either N or N-1 can be specified on the CRITERIA subcommand). STANDARDIZED is the default distance function, with sum of squares = N.

PROXIMITIES
The data contains proximity and represents a square matrix of measures between objects. With this keyword, the variable list on the SOURCEID subcommands specify a proximity matrix directly.

Example

GET FILE ’states.sav’.
PROXMAP
 /SOURCEID DATA =  popul income illit life homic school freeze fail
           NAME = $SRC1 .

In this example, the multivariate data (default) concern demographic and quality of life variables on the 50 US states. Standardized Euclidean distances are derived.

The single source gets the name $SR1 and label SOURCE 001 (default)