Re: [ontolog-forum] Constructs, primitives, terms

[Amanda Vizedom <amanda.vizedom@xxxxxxxxx>]

David,

On Fri, Mar 2, 2012 at 10:52, David Eddy <deddy@xxxxxxxxxxxxx> wrote:

Amanda -

On Mar 2, 2012, at 10:04 AM, Amanda Vizedom wrote:

For this kind of usage, better accuracy and intuitiveness come from calling a concept ont1:0395 (some existing label), and deciding based on meaning whether it should also have labels ("gruefulit", sysA) or ("gruefulit", sysB) or neither. 

Surely there's a rollicking good story behind "gruefulit"?

Not really. I wanted to use a nonsense word to make the point general. "Grue" has a historical use in certain philosophical discussion, introduced by Nelson Goodman, and popped it to my head, but of course that means that "grue" is already meaning-laden to many folks with philosophical training, and I didn't want to set any those off, either. ;-)  So I just extended it in an arbitrary way that I was pretty confident wouldn't mean anything, but looks and sounds reasonably like it could be a word. 

 

I think I agree with this approach... are you saying that by some process gruefulit A & B are presented/made known to a human—I seriously doubt if I'd trust an automated machine translation process—and the human decides equivalence or not?

Do I grok that correctly?

Yes, that's what I've been focusing on in talking about mapping. So far, most mapping is done manually or manually with some automated assistance (e.g., suggestions). 

I should add, though, that there are techniques for automated text analysis that use the labels on a concept to pick up references to it, making use of NLP standards like stemming, morphing, etc., and using various statistical techniques. Some folks have turned these to mapping and ontology evaluation tasks (see work by Kevin B. Cohen, et al, for example).  Some of these are pretty impressive. 

Even better, from what I've seen, are some techniques that use both statistical processing of expressions and lexical info and select ontological relations between concepts mapped to those expressions to perform disambiguation. These techniques could also be applied to mapping, if they haven't been already. Actually, I'd bet that they have been so applied and I just haven't (knowingly) encountered that work yet. 

Amanda

 

___________________

David Eddy

deddy@xxxxxxxxxxxxx

781-455-0949

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
 

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (01)