SAS Text Miner 3.1 contains a new procedure, PROC DOCPARSE, to parse the
document collection.  There are a few PROC statement experimental
options for the OUT= data set that are not in the procedure
documentation.  The experimental options, ADDTERM, ADDTAG , ADDPARENT,
and ADDOFFSET, add the corresponding columns to the OUT= data set:

  _TERM_   - the term.
  _TAG_    - the tag of the term.
  _PARENT_ - the parent of the term.
  _OFFSET_ - the order of the term in the order in which it was returned
             from the natural language processor to PROC DOCPARSE.

In the following example, PROC DOCPARSE finds all maximal groups,
entities, and parts of speech for a single document.

data cars;
   input text $1-70;
   datalines;
I have a white pearl Honda Insight.
Mary has a silver Volkswagen Beetle.
Johnny wants a red Honda Civic Hybrid.
;
run;

proc docparse
   data=work.cars
   entities=yes
   stemming=yes
   tagging=yes
   ng=max
   key=dropkey
   out=dropout addterm addtag addparent addoffset;
   var text;
run;

Operating System and Release Information