Dear Harish,
In my point of view, we can't compare similarity measures by efficiency since
they benefit from different methods. for example Lin proposed commonality
between two words as the information content, Jiang and Conraths combined two
approaches: information content of the two nodes as well that of their
mscs(most specific common subsumer) while Leacock and Chodorow proposed the
normalized path length.
Based on the earlier evaluation these three measures are more similar to
human rating similarity. So choosing between these measures is strongly depends
on your application and methodology. I recommend you to have a look at the
theory of these measures to have a clearer idea.
Best Regards,
Yashar Mehdad.
Harish Garapati <[EMAIL PROTECTED]> wrote:
Hi,
I am working on recognising entailment between two sentences and started
using wordnet similarity tool,for finding out the similarity measure between
two words.I am new to this tool and currently using "Lin" measure to find the
semantic similarity between two words. Is there any measure that would better
represent the similarity between two words?
I found out that there are cases where the Lin measure doesnot give the
semantic measure as expected.
Thanking you,
Harish.
---------------------------------
Forgot the famous last words? Access your message archive online. Click here.
---------------------------------
Be a better friend, newshound, and know-it-all with Yahoo! Mobile. Try it now.
