W3C home > Mailing lists > Public > semantic-web@w3.org > May 2005

Re: WordNet in RDF

From: Jeen Broekstra <jeen@aduna.biz>
Date: Tue, 31 May 2005 10:19:04 +0200
Message-ID: <429C1DF8.6020601@aduna.biz>
To: Alvaro Graves <agraves@dcc.uchile.cl>
Cc: semantic-web@w3.org

Alvaro Graves wrote:
> 
> Hi, my name is Alvaro Graves, I'm working on my MS in the University of 
> Chile. My thesis consists in a RDF Representation of WordNet and to 
> create some application over it.
> 
> Currently I have parsed almost the entire wordnet 2.0 relations to RDF. 
> The schema I created was based in one you had in 
> http://www.w3.org/2001/sw/BestPractices/WNET/wordnet-sw-20040713.html
> 
> I would appreciate a lot if you could check  my schema and files of 
> wordnet and give me your comments.
> 
> The URI is http://www.dcc.uchile.cl/~agraves/wordnet/index.html

First of all: nice effort!

I'll have a closer look at it later, but I just noticed one thing 
right off that you might want to change.

In your schema, you are defining the hyponym property with multiple 
domains (VerbSynSet and NounSynSet). I think this is not really a good 
idea: multiple domains/ranges are defined as an intersection in RDFS. 
The result would therefore be that every concept that has a hyponym 
property would be classified as _both_ a VerbSynSet and a NounSynSet.

Two easy solutions:

  1. use a single domain that is the lowest common ancestor of the two
     classes, for example 'SynSet'.
  2. create domain-specific subproperties.

I tend to favor the first option because introducing a lot of 
subproperties can clutter up your schema.

Jeen
-- 
Jeen Broekstra          Aduna BV
Knowledge Engineer      Julianaplein 14b, 3817 CS Amersfoort
http://aduna.biz        The Netherlands
tel. +31 33 46599877
Received on Tuesday, 31 May 2005 08:16:38 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:05 GMT