Daniel, thanks for your example, it looks very good!
On Mon, Jun 1, 2015 at 3:31 PM, Daniel Kinzler daniel.kinzler@wikimedia.de wrote:
If I understand you correctly, your are looking for a way to describe different meanings or facettes of a *word*, building clusters based on what other concept each of these meanings is related to.
Since Wikidata does not (yet) deal with words at all, we can only defer this until we do (see the Wikidata/Wiktionary proposal). But others have done this: have a look at the "Wortschatz" project by the University of Leipzig:
http://corpora2.informatik.uni-leipzig.de/res.php?corpusId=deu_newscrawl_201...
The graph at the bottom shows clusters of collocations for each meaning/facette. It's a bit hard to find good examples though, usually one meaning is very dominant.
Am 01.06.2015 um 15:14 schrieb David Cuenca Tudela:
Hi Leila
On Mon, Jun 1, 2015 at 9:32 AM, Leila Zia <leila@wikimedia.org mailto:leila@wikimedia.org> wrote:
Please correct me if I'm wrong: it seems to be tne case that thesequestions
are more about the interfaces and technologies built on top ofWikidata, not
so much about Wikidata itself (except for the point that thefundamental
data should exist in Wikidata).Yes, in a way it is more about representation that anything else, but it
also
has to do with the conceptual framework. In the current organization it
is
assumed that I want to know about an specific Q item, whereas in my
thinking
about signals I would prefer to have an overview of all items that are
related
to a keyword.
For instance if I enter Chopin, either I have to select one item from a
list or
I have to perform a search, there is no middle way of displaying an
overview of
all items grouped by the kind of relationship that they might have to the keyword chopin. In a way it is a bit like creating a disambiguation page
on the
fly, with the added difficulty of grouping elements that belong
together. For
instance if I search Bach, it would make sense to group people with the
string
"bach" related to the same family together, and divide it by topic, like
a sort
of disambiguation page for data: https://en.wikipedia.org/wiki/Bach_%28disambiguation%29
What is difficult is to find an automatic arrangement that works for most situations, or explore a different way of creating data disambiguation
pages,
perhaps based on current disambiguation items. Is there any way to make
this
item more useful with some visualization of items it disambiguates? https://www.wikidata.org/wiki/Q107809
Cheers, Micru
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
-- Daniel Kinzler Senior Software Developer
Wikimedia Deutschland Gesellschaft zur Förderung Freien Wissens e.V.
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata