PROXIMITIES writes a variety of proximity matrices, each matrix with ROWTYPE_ values of PROX. PROXIMITIES neither reads nor writes additional statistics
with its matrix materials. See Format of the Matrix Data File (PROXIMITIES command) for a description
of the file.
The matrices that PROXIMITIES writes can be used by PROXIMITIES or other procedures. Procedures CLUSTER and ALSCAL can read a proximity
matrix directly. Procedure FACTOR can read a correlation matrix that is written by PROXIMITIES, but RECODE must first be used to change the ROWTYPE_ value PROX to ROWTYPE_ value CORR. Also, the ID subcommand cannot be used on PROXIMITIES if the matrix will be used in FACTOR.
If VIEW=VARIABLE, the variables in the matrix file will have the names and labels
of the original variables.
If VIEW=CASE (the default), the variables in the matrix file are named VAR1, VAR2, ...VARn, where n is the
sequential number of the variable in the new file. The numeric suffix n is consecutive and does not necessarily
match the number of the actual case. If there are no split files,
the case number appears in the variable label in the form CASE m. The numeric suffix m is the actual case number and may not
be consecutive (for example, if cases were selected before PROXIMITIES was executed).
If VIEW=CASE, a numeric variable CASENO_ is
added to the matrix file. Values of CASENO_ are the case numbers in the original file.
The new file preserves the names and values of any
split-file variables that are in effect. When split-file processing
is in effect, no labels are generated for variables in the new file.
The actual case number is retained by the variable ID.
Any documents that are contained in the active dataset
are not transferred to the matrix file.