[Attempto] generating OWL ontology

[Kaarel Kaljurand]

Hi,

On Mon, Mar 30, 2009 at 4:54 PM, Alireza <alireza.khoshkbari at gmail.com> wrote:
> As ACE lexicon has been predefined, unknown words should be defined.
> therefore,unknown words can be prefixed by their word class in the ACE text.
> Use n: for nouns, p: for proper nouns, v: for verbs, and a: for adjectives
> and adverbs.
> (Source:http://attempto.ifi.uzh.ch/site/docs/ace_troubleshooting.html)
>
> Alireza.
>
> On Mon, Mar 30, 2009 at 7:03 PM, David Whitten <david.whitten.cyc at gmail.com>
> wrote:
>>
>> I'm not familiar with the idea of using a letter prefix with a colon
>> before
>> an English word.  Does this make the OWL information generated work
>> better? Is this supposed to be an RDF namespace ?

The ACE parser that we offer (APE) has a built-in lexicon that knows
100,000 word forms. To parse sentences that use word forms not in
the lexicon, one can prefix the word forms with word class markers
(n, v, p, etc.) or rely on unknown word guessing. This feature was once added
for debugging purposes and is not very general, e.g. the lemma information
that is contained in the lexicon cannot be given this way.

So I would recommend using the user lexicon instead of prefixes.
AceWiki and ACE View generate the user lexicon when new words
are introduced and provide a front-end for editing the wordforms
that correspond to the lemma (i.e. OWL entity in case of AceWiki and ACE View).

If you use APE directly (i.e. the webservice of APE), then you can
pass the Prolog-formatted user lexicon to APE either in a file (ulexfile)
or in a string (ulextext), see:

http://attempto.ifi.uzh.ch/site/docs/ape_webservice.html

--
kaarel