( Correction: the "Vector Space..." article is not actually directly related to the OP's problem but does deal w/ the "closeness of words" concept. In that regard, below is the corrected version of my earlier reply.
UPDATE (Mar 3 2003): If anybody is still interested, i have rounded up some relevant things under "string munging". )
See Vector Space Search Engine article which does some similar things.
The String::Similarity and String::Approx modules may also be of interest. Below are the descriptions (from FreeBSD ports)...
In reply to Re: Calculating "similarity"
by parv
in thread Calculating "similarity"
by Anonymous Monk
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |