Re: ontolex telco this week

Dear Philipp, All

last week it emerged the need for a few other examples regarding the
metadata module. In particular, we were asked to show:

   1. how to avoid the introduction of a separate void:Dataset for a
   lexicon or an ontology;
   2. how to use an existing Lexicon.

Unfortunately, we were busy with other duties, so we had not the
possibility to elaborate concrete examples in time for today's telco.
Nonetheless, I put below some thoughts highlighting what the examples will
possibly look like.

Concerning the second request, we want to use WordNet to lexicalize the
FOAF ontology (which is the reference dataset for the already provided
examples). Furthermore, we believe that the use of WordNet should be
demonstrated through two examples, depending on whether ontology entities
are linked to lexical entries or synsets.

The former scenario is very similar to the already provided example, in
which we lexicalize FOAF with respect to a dedicated lexicon. Conversely,
the latter scenario is different, since we are connecting ontology elements
to lexical concepts rather than lexical entries.

Accordingly, in the original proposal [1] we introduced a dedicated class,
named lime:LexicalLinkSet. With respect to this scenario, we should
certainly describe how well synsets cover the ontology. At the same time,
we should discuss whether we desire to also provide statistics in terms of
plain lexical entries.

As an example, suppose that we associate the class foaf:Person with the
synset {individual, mortal, person, somebody, someone, soul} (
http://wordnet-rdf.princeton.edu/wn31/100007846-n).

This association counts as a single link from the viewpoint of synsets.
Then, we could compute metrics such as the percentage of ontology entities
that are associated with any synset.

However, we could be interested as well in the fact that the aforementioned
link actually contributes 6 lexical entries. We could then compute the
average number of lexical entries for an ontology concept.

[1]
https://github.com/jmccrae/ontolex/blob/ed9ee2a214b26780c64049e18827c36e72bb6a50/Ontologies/lime.owl#L80



2014-10-29 8:17 GMT+01:00 Philipp Cimiano <cimiano@cit-ec.uni-bielefeld.de>:

> Dear all,
>
>      I will not be available for the ontolex telco this week, but I
> suggest that John/Paul lead the telco this week to continue the technical
> discussion on:
>
> 1) metadata examples from Armando and Manuel. I was fine with them. Others
> should comment as well.
> 2) taxonomy of Variants proposed by Elena. Please discuss this again to
> reach consensus (see proposal below)
> 3) example of Jorge on how to model provenance from translations
>
> I would be very grateful if someone could take minutes.
>
> The minutes from last telco are here:
>
> http://www.w3.org/2014/10/24-ontolex-minutes.html
>
> The minutes are quite sparse, though ;-)
>
> So let me say that the proposal is to have an uper class Variant with
> subclasses FormVariant, MorphologicalVariant (instead of LexicalVariant),
> SenseVariant, with TerminologicalVariant and Translation as subclasses of
> the latter. Please see if there is agreement on this point in the ontolex
> community.
>
> Kind regards,
>
> Philipp.
>
> --
> --
> Prof. Dr. Philipp Cimiano
> AG Semantic Computing
> Exzellenzcluster für Cognitive Interaction Technology (CITEC)
> Universität Bielefeld
>
> Tel: +49 521 106 12249
> Fax: +49 521 106 6560
> Mail: cimiano@cit-ec.uni-bielefeld.de
>
> Office CITEC-2.307
> Universitätsstr. 21-25
> 33615 Bielefeld, NRW
> Germany
>
>
>


-- 
Manuel Fiorelli