oops - if species concepts are classes, then I guess I should replace each
occurence of owl:sameAs (below), with owl:equivalentClass ...
On Fri, 4 Jun 2010, joel sachs wrote:
Pete,
I notice that you use skos:closeMatch (instead of owl:sameAs) to assert near
equivalence amongst species concepts. E.g.,
<owl:Class rdf:about="http://lod.taxonconcept.org/ses/mCcSp#Species";>
<skos:closeMatch
rdf:resource="http://dbpedia.org/resource/Monarch_(butterfly)"/>
In an area as tricky as species concepts, I agree that skos:closeMatch has
advantages over owl:sameAs. (For example, closeMatch is not transitive, which
makes it harder for errors to propagate.)
You then make a series of foaf:isPrimaryTopicOf statements, such as
<owl:Class rdf:about="http://lod.taxonconcept.org/ses/mCcSp#Species";>
<foaf:isPrimaryTopicOf
rdf:resource="http://en.wikipedia.org/wiki/Monarch_(butterfly)"/>
But foaf:isPrimaryTopicOf is an inverse functional property. Since I assume
you agree with the assertion (from dbpedia) that
<rdf:Description
rdf:about="http://en.wikipedia.org/wiki/Monarch_(butterfly)">
<foaf:primaryTopic
rdf:resource="http://dbpedia.org/resource/Monarch_(butterfly)"/>
you are, in fact, asserting that your Monarch Butterfly and dbpedia's Monarch
Butterfly are owl:sameAs each other. (Is there a reason you need
foaf:primaryTopic, instead of, say, dc:subject?)
I look forward to using this dataset -
Regards,
Joel.
On Fri, 4 Jun 2010, Peter DeVries wrote:
Hi Everyone,
I am working on a new model for species concepts at TaxonConcept.org.
These are an extension of my work on GeoSpecies, and are also designed to
investigate mapping names to concepts as part of the Global Names
Initiative<http://gni.globalnames.org/>
.
The main difference between these and GeoSpecies are that each species
concept is a class.
This is to address some users who feel that species are best modeled as a
class.
Other groups would like species modeled as an instance.
This is mainly those groups that are dealing with large numbers of species
occurrence records.
To deal with these different needs I have come up with the following
solution.
Each species concept has related "Tags" that are instances.
For example
In this RDF there is a description of the following species concept
http://lod.taxonconcept.org/ses/mCcSp.rdf
there is a description of the following species concept
http://lod.taxonconcept.org/ses/mCcSp#Species
URIBurner <
http://linkeddata.uriburner.com/about/html/http/lod.taxonconcept.org/ses/mCcSp%01Species
<txn:speciesConceptHasSpeciesTopicTag rdf:resource="
http://lod.taxonconcept.org/ses/mCcSp#Topic"/>
<txn:speciesConceptHasSpeciesOccurrenceTag rdf:resource="
http://lod.taxonconcept.org/ses/mCcSp#Occurrence"/>
<txn:speciesConceptHasSpeciesIndividualTag rdf:resource="
http://lod.taxonconcept.org/ses/mCcSp#Individual"/>
I created the #Topic tag because I originally was thinking about using that
to label those resources that have that species as a topic.
Some think this may be unnecessary, but I have left it in for now while I
think about it.
The "mCcSp#Occurrence" tag is to indicate that the entity is an occurrence
of that species.
These are instances of the class txn:SpeciesOccurrenceTag, which allows
someone to partition the data set into only the information that is about
species occurrence records.
The "mCcSp#Individual" tag is to represent individuals of that species.
Another feature of the species concept RDF is that it maps the species
concept to related entities on the LOD and as foreign keys in other data
sets.
The data set is available via my semantic site map at:
http://lod.taxonconcept.org/sitemap.xml.gz
And is currently live on the LOD. A SPARQL Endpoint is being setup.
The data set contains about 86,000 species concepts.
** We have notice that there maybe some problems with the Fungi section, so
I would avoid those for the time being. **
I have created a smaller example RDF dump file at:
http://lod.taxonconcept.org/taxonconcept_subset.rdf.gz
This contains a combination of species concepts and related linked data
including Image galleries and a small sample of occurrence records.
Here are some examples as viewed through the Sigma service.
Monarch Butterfly Danaus plexippus <
http://sig.ma/search?pid=5d903f44eb963350b77074b7d9e80ff9 >
Eastern Tree Hole Mosquito Ochlerotatus triseriatus <
http://sig.ma/search?pid=5d903f44eb963350b77074b7d9e80ff9 >
Additional Examples can be found at:
http://www.taxonconcept.org/example-taxa/
I also have some example SPARQL Queries that work on the LOD.
http://www.taxonconcept.org/example-sparql-queries/
I would appreciate feedback on these models and any suggestions for how
they
could be improved. :-)
Thank you for your time and consideration.
- Pete
Ontology: http://lod.taxonconcept.org/ontology/txn.owl
Ontology Doc: http://lod.taxonconcept.org/ontology/doc/index.html
----------------------------------------------------------------
Pete DeVries
Department of Entomology
University of Wisconsin - Madison
445 Russell Laboratories
1630 Linden Drive
Madison, WI 53706
GeoSpecies Knowledge Base
About the GeoSpecies Knowledge Base <http://about.geospecies.org/>
------------------------------------------------------------
