BIB-VERSION:: CS-TR-v2.0
ID:: UCB//S2K-91-10 
ENTRY:: February 14, 1994 
TITLE:: A Method for Refining Automatically-Discovered
        Lexical Relations: Combining Weak Techniques for
        Stronger Results 
DATE:: 
AUTHOR:: Grefenstette, Gregory 
AUTHOR:: Hearst, Marti A.  
PAGES:: 9 
ABSTRACT:: Knowledge-poor corpus-based approaches to natural
language processing are attractive in that they do not incur
the diffiulties associated with complex knowledge bases and
real-world inferences. However,  these kinds of language pro-
cessing techniques in isolation often of not suffice for a 
particular task; for this reason we are interested in finding
ways to combine various techniques and improve their results.

Accordingly, we conducted experiments to refine the results
of an automatic lexical discovery technique by making use of
a statistically-based syntactic similarity measure. The dis-
covery program uses lexico-syntactic patterns to find instances
of the hyponmy relation in large text bases. Once relations of
this sort are found, they should be inserted into an existing
lexicon or thesaurus. However, the terms in the relaiton may
have multiple senses, thus hampering automtic placement. In 
order to address this problem we tried to make a term-similarity
determination technique choose where, in an existing thesaurus,
to install a lexical relation. The union of these two corpus-
based methods is promising, although only partially successful
in the experiments run so far. Here we report some prelimimary
results, and make suggestions for how to improve the technique
in future.
RETRIEVAL:: postscript (in all.ps)
END:: UCB//S2K-91-10