torm -a -o N /Data/tmp/t9 (or slim,twoyear) < infile >outfile
-a specifies associations (requires t9 database)
-o specifies number of nearest neighbors to obtain
N = number of neighbors desired
-t 1 or 2 (default to 1) for # of processors
/Data/tmp/t9 (or slim) t9 is full file; slim is w/o associations
< infile just a list of words
> outfile what program provides
torm -a -o /Data/tmp/t9 (or slim) kidmat
then type in word(s) and ^d to output to screen
Output looks as follows:
neighbors of: cat
cat = 0.0000 5.01 1036
man = 474.1234 4.96 49
1 2 3 4
Column 2 is semantic distance in RSUs.
Column 3 is mean number words cooccuring between two items within
10 word window.
Column 4 is raw frequency of two items cooccurring.