I'm trying to improve a research paper I did on measures in ontologies by getting a correlation of a certain measures results in WordNet to human similarity scores. What I'll compare that to is the list of correlations to those same human similarity scores in a paper by Nuno Seco, Tony Veale and Jer Hayes called "An Intrinsic Information Content Metric for Semantic Similarity in WordNet"
http://eden.dei.uc.pt/~nseco/ecai2004b.pdf

My problem is that I don't know a thing about WordNet, I only know it can be thought of as an ontology. What I need in order to apply this new measure is to be able to find out for any word in WordNet what words subsume it and which it subsumes. There are other ways to phrase that, like what are the hyponyms of a given word, I don't really know a non-technical way to describe it. Basically, I know that "a nickel is-a coin" and "nickel is-a metal" and that those two senses of the word "nickel" are in the same synset but I don't know how to get WordNet to tell me that " 'coin' subsumes 'nickel' ," that if you drew it as a graph coin would have an edge pointing down to nickel among many others.

2/25/2007 10:41:05 PM

man whats all that moonman talk

2/25/2007 11:08:42 PM

yeah, sorry. I'm kinda grasping at straws, I know just enough to be dangerous in that I don't even know enough to phrase it non-moonman.
The easiest way to rephrase might be to talk about how Cliff Joslyn casts ontolgies as "directed acyclic graphs" This one I can translate to English: WordNet is a bunch of words that are related in only one direction, with a root at the top and all edges directed downwards representing that each lower concept is more specific, as 'dime' is a more specific subset of the concept 'coin'
My problem is that I can't get WordNet in that form, I can look words up and get definitions and synonyms but not which words 'contain' which others.

2/26/2007 1:23:50 PM

[Edited on February 26, 2007 at 10:59 PM. Reason : ]

2/26/2007 10:55:14 PM

?Donde estan los otros 'Moonmen'?

http://eden.dei.uc.pt/~nseco/ecai2004b.pdf

[Edited on March 9, 2007 at 8:40 PM. Reason : mmkay?]

3/9/2007 8:12:19 PM