I've been silent for a while so I'll throw my $0.02 too. (01)
For me, an ontology can induce a taxonomy if it is sufficiently well
axiomatized
for a reasoning system to relate axioms about asserted relationships
(e.g., subsumption)
with axioms about asserted properties (e.g., an attribute value or a
semantic annotation like
those made with OntoClean's sign language "+/-/~" and alphabet soup
"I,M,O,U,S,...") (02)
A classifier combines these two kinds of assertions (i.e., relationships
& properties) to infer
additional relationship & property facts and/or additional lemmas,
theorems, ... until one of two things happen: (03)
a) the classifier reaches a (stable) fixed point where no new
relationship/property can be logically infered
b) the set of infered & asserted
relations/properties/lemmas/theorems/axioms....
is logically inconsistent with the logical axioms about kinds/types of
relations & properties we use as ground truth. (04)
In OntoClean, there are multiple examples of logial axioms that relate
OntoCLean's properties (e..g, +R, ~R)
with relationships (e.g., T1 subsumes T2) to yield useful consistency
axioms: (05)
a concept/relation T1 annotated ~R cannot subsume a concept/relation T2
annotated +R (06)
Unfortunately, many taxonomies are defined with just "assertions" and
are very weak in terms of inferable relations/properties.
The worst kind of taxonomy are those where the axiomatic basis that
would have produced, by classification, the taxonomy as
defined is implicitly defined (e.g., in someone's head). In that sense,
you could say that combining taxonomies and ontologies
requires: (07)
- a context of classification that identifies:
-- a "ground" ontology of concepts, relations, properties
-- a "classification" ontology of annotations, values, etc... that we
can use to classify elements of the ground ontology
-- an annotation that tags relations/properties (e.g., "dog") from the
ground ontology with individuals in the classification ontology (eg., +I)
-- logical axioms that map logical properties from the ground ontology
(e.g,. concept subsumption) to logical properties in the classification
ontology (e.g., ~R cannot subsume +R)
-- the type of logic required to ensure that inferences made using the
logical axioms are sound and consistent (this might require annotations
about the transitivity of some properties and corresponding logical
axioms that enable the classifier to use these annotations -- e.g.,
transitivity) (08)
- once the context is properly identified, then we should have something
more or less equivalent to a "classification service"
and match that with a tool / implementation of a classifier that is
capable of providing that service (e.g., Racer, Pellet, ...)
within the context of the logic used for specifying, reasoning and
describing logical statements about classifying the elements
of the ground ontology into a hierarchical structure according to the
asserted and infered classification relationships & properties.
If logically sound, then the resulting hierarchical structure would be
what I'd call a taxonomy. (09)
This is a quick note, I may have got some details wrong but I think the
gist of it is there based
on earlier presentations about "classification" and "taxonomies" from
the OntoClean folks and
Alan Rector's presentations a couple of years ago at the Protege conference. (010)
It may be a good time for me to put on my fire jacket... (011)
-- Nicolas. (012)
Jayne E Dutra wrote: (013)
> Hello All,
>
> I am re-posting this question since it seems there was some anomaly
> with the ontolog server. I hope someone out there can help!
>
>
>**********************************************************************************
>
>
> Hello All,
>
> I am looking for material regarding the relationship between ontologies
> and taxonomies. I have seen a couple of slides showing taxonomies as a
> low
> level or foundation level layer in some presentations regarding
> ontologies, but it seems to me to be given short shrift. After all, if we
> are to implement the Semantic Web in a robust way, we all need to
> agree on
> the meaning of specific terms that may be used in ontologies or else the
> higher level logic becomes a lot less meaningful.
>
> If anyone knows of papers, articles, presentations, etc, on this topic,
> please let me know.
>
> Thanks very much in advance,
>
> Jayne
>
>
>
>
> *******************************
> Jayne Dutra
> Jet Propulsion Laboratory
> Information Architecture and Semantic Engineering
> JPL Knowledge Management
> 818-354-6948
> ******************************
>
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Subscribe/Unsubscribe/Config:
> http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/ To Post:
> mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>
> (014)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Unsubscribe/Config:
http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (015)
|