|
Excellent, Aldo! I do recall reading your papers earlier,
and think they are indeed very useful for our effort here.
Folks, I'd suggest looking at these. For example, the first
paper below discusses some principles and parameters for measuring ontology
goodness (p. 10-12):
Transparency (explicitness of organizing principles)
• Computational integrity and
efficiency
• Meta-level
integrity
• Flexibility
(context-boundedness)
• Compliance to
expertise
• Compliance to procedures
for extension, integration, adaptation, etc.
• Generic accessibility
(computational as well as commercial)
• Organizational
fitness
These
are described in some detail, and will assist us in our
effort.
Thanks,
Leo
Hi Leo, this seems a much more concrete thread than defining "ontology"
:).
Regarding dimensions/aspects, you may want to check the work made by my
group on formalizing qualitative criteria for ontologies. That is a quite
comprehensive framework, and you can take advantage of it during these early
attempt at singling out dimensions.
The papers and the OWL ontology referenced below describe ontology
dimensions, measures, and qualitative principles. We have modelled them by
using extensive reification on top of DOLCE and the Descriptions and
Situations reification vocabulary:
- Gangemi, Catenacci, Ciaramita, Lehmann: Modelling Ontology Evaluation
and Validation, ESWC 2006:
http://www.loa-cnr.it/Papers/ESWC_GangemiEtAl_Final2.pdf.
- Gangemi, Catenacci, Ciaramita, Lehmann: Qood grid: A metaontology-based
framework for ontology evaluation and selection, EON 2006:
http://km.aifb.uni-karlsruhe.de/ws/eon2006/eon2006gangemietal.pdf.
- An OWL ontology for the metamodel:
http://www.loa-cnr.it/ontologies/EVAL/oQual.owl.
Best
Aldo
At 13:28 -0600 29-01-2007, AJ Chen wrote:
Content-type:
text/html
Content-disposition: inline
Content-description: HTML
Version of Message
Leo, thanks for putting a list of dimensions
together. As I suggested earlier, it would be helpful to the discussion if
we could drill down these dimensions with examples of existing ontologies or
ontologies just made up for the discussion.
Working from application development angle,
I look at relevant ontologies mainly in two dimensions: (1) the application
areas or domains an ontology is design for; and (2) the level of granularity
an ontology has. The main purpose of surveying existing ontologies in such a
way is for re-using of the existing ontological classes and properties and
for increasing the chance for interoperability. And so, my hope here is that
we can brainstorm on any possible way that will make the task of comparing
ontologies and their classes and properties easier for application
developers. More importantly, a clear framework for ontology comparison will
potentially help people produce new ontologies that are easier to be reused
and integrated.
The examples I'm looking at are dublin core,
FOAF, vCard, FuGO, EXPO. The questions are:
(1) how do we categorize them from
application and granularity dimensions?
(2) can we identify/develop a common
framework that improve our understanding of these different ontologies and
thus increase reuse and integration?
(2) what suggestions can we make to the
owners of the ontologies that will increase the re-usability and
interoperability of their new versions of ontologies?
Any new thoughts or comments?
AJ
Quoting "Obrst, Leo J."
<lobrst@xxxxxxxxx>:
> Folks,
>
> To begin to
address the "typing" of ontology that the Ontology Summit
>
prospectively would like to address, I'm beginning this new thread.
I'd
> suggest using the subject line to clearly distinguish any NEW
topic, so
> that skimmers/samplers/surfers can focus more
meaningfully. Why?
> Because we really don't yet have
metadata-annotated or "semantically"
> threaded email -- except for
subject lines. For example, currently we
> have the topic "ontology as
logical theory?".
>
> I'm bringing forward my suggestion
(message dated Thursday, January 25,
> 2007 2:14 PM, under the Subject
= Re: [ontology-summit] Defining
> "ontology"), simply as a strawman,
slightly enhanced with a canonical
> dimension/feature name (e.g.,
Formality) and a scale of values (e.g.,
> Informal=0, Formal=1). I
don't know if this will prove useful, but let
> us know your thoughts.
Note that these are not pejorative: the feature
> name and the values
are relatively arbitrarily chosen (though I have my
> own
biases).
>
> Please comment on/modify/trash these. Note if you
strongly believe in
> one definition or think this effort is not
worthwhile or the dimensions
> are mischaracterized, please let us all
know.
>
>
> Dimensions of Ontology Types:
>
>
1) Formality: Informal (Formality = 0) vs. Formal (Formality =
1)
> 2) Expressivity: Expressivity of the semantic model (i.e.,
underlying
> knowledge
> representation language or logic) [No
scale determined yet]
> 3) Concept-based: Term (Concept-based = 0) vs.
concept (real world
> referent)(Concept-based = 1)
> 4)
Mathematical Ordering: Mathematical ordering, structure, definition
>
of the privileged
> parent-child relation: [No scale determined
yet]
> 5) Application focus/use cases, etc. (part of this is precision
of the
> service needed, e.g., metadata/topic terms for a document to
aid in
> broad doc topic retrieval vs. a semantic service query,
specfication,
> or composition): [No scale determined
yet]
> 6) Granularity (precision, scope): [No scale determined
yet]
> 7) Development Philosophy: Empirical (bottom-up) [0] vs.
Rationalist
> (top-down) development [1]
> methodology (i.e.,
arbitrary folks add or annotate terms/concepts vs. a
> rigorous
ontology development) [No scale determined yet; Some
> combination?
Middle-out? But what does that mean?]
> 8) Human-Coded: Human-coded
[1] vs. machine-learned/generated [0]
> 9) Automated reasoning (and
complexity of that, i.e., one could have
> transitive closure or
subsumption down a subclass graph vs.
> theorem-proving): [No scale
determined yet]
> 10) Descriptive vs. prescriptive (i.e., a
commonsense or
> conceptually-profligate ontology vs. an ontology that
specifies that
> this is the way the world is): [No scale determined
yet]
> ...
>
> Other criteria perhaps address properties
of the content, i.e., Average
> density/bushiness: [Probably a real
scale, once we define density,
> bushiness],
etc.
>
>
> Some notes:
> 1. Expressivity,
Mathematical Ordering, and Automated Reasoning are
> probably
related.
> 2. Application focus/use cases: this is still nearly
arbitrary, but I
> would like us to think about recall/precision of
the application/use
> case as perhaps a better scale or paired scales,
i.e., [Recall = 0,
> Precision = 0] vs. [Recall = 1, Precision = 1],
but then we need to
> define Recall and Precision. Or there may be
some normalized score that
> factors both Recall and Precision into a
composite real number. Then
> again, a better dimension may be
application complexity if that can be
> agreed on, i.e., Constant,
Linear, Logorithmic, Exponential, ...
> NP-Complete, ... Of course, if
we go with this latter notion, then it
> is probably closely related
to (1) above, i.e., if a problem is
> characterizable as NP-Complete,
then there is probably a corresponding
> expressivity model
requirement. This gets into expressive complexity,
> i.e., the
correlation between the expressiveness of the language and
>
complexity of the computation.
>
> Thanks,
> Leo
>
_____________________________________________
> Dr. Leo
Obrst The MITRE Corporation, Information
Semantics
> lobrst@xxxxxxxxx Center for Innovative
Computing & Informatics
> Voice: 703-983-6770 7515 Colshire Drive,
M/S H305
> Fax: 703-983-1379 McLean, VA 22102-7508,
USA
>
>
>
_________________________________________________________________
>
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
>
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
>
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
> Community
Files: http://ontolog.cim3.net/file/work/OntologySummit2007/
>
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2007
>
Community Portal: http://ontolog.cim3.net/
>
--
AJ
Chen, PhD
http://www.web2express.org
"Open
Data on Semantic Web"
_________________________________________________________________
Msg
Archives:
http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config:
http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe:
mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files:
http://ontolog.cim3.net/file/work/OntologySummit2007/
Community Wiki:
http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2007
Community
Portal: http://ontolog.cim3.net/
--
Aldo Gangemi
Senior Researcher
Laboratory for Applied
Ontology
Institute for Cognitive Sciences and Technology
National
Research Council (ISTC-CNR)
Via Nomentana 56, 00161, Roma, Italy
Tel:
+390644161535
Fax: +390644161513
aldo.gangemi@xxxxxxxxxxx
http://www.loa-cnr.it/gangemi.html
icq# 108370336
skype
aldogangemi
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2007/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2007
Community Portal: http://ontolog.cim3.net/ (01)
|