W3C home > Mailing lists > Public > public-lod@w3.org > February 2012

Re: How to align ontology terms that are similar but are set to potentially incompatible domain?

From: François Scharffe <francois.scharffe@lirmm.fr>
Date: Fri, 10 Feb 2012 14:37:03 +0100
Message-ID: <4F351D7F.2040006@lirmm.fr>
To: Peter DeVries <pete.devries@gmail.com>
CC: public-lod@w3.org
Hi Pete,

This is an interesting use case showing the complexity of aligning 
ontologies, and the lack of a standard way to represent alignments.

I couldn't access http://lod.taxonconcept.org/ so I'll give you an 
examples using the Biological Taxonomy Vocabulary (biol) the Wildlife 
Ontology (wo) and GeoSpecies (gs).

The easiest case is term equivalence, as it seems to be between
wo:scientificName
and biol:name
You can use owl:equivalentProperty between them.

It becomes more complex if the terms are not fullly equivalent:

on one side gs:hasScientificName
has domain gs:SpeciesConcept
which is subclass of gs:TaxonConcept

on the other side wo:scientificName
has domain wo:TaxonName which is related to wo:TaxonRank through the 
wo:name object property, and wo:Species is subclass of wo:TaxonRank

so it seems that gs:hasScientificName
applies to a narrower domain.

You can thus use skos:narrower to indicate this relation.
It is more precise than skos:closeMatch but still quite vague.

So it would be better to specify the exact relation between the two 
properties, by specifying the type of the wo instances should be wo:Species.
In RIF BLD:
Forall ?x ?y  ( gs:hasScientificName(?x ?y) :- and( wo:scientificName(?x 
?y) rdf:type(?x, wo:Species) ) )


Now the problems:
- The skos approach raise problems given the semantics of the skos terms,
the properties aligned using skos:closeMatch or skos:narrower become 
skos Concepts. This might lead to some unwanted effects when using a 
reasoner.
- The RIF rule does not completely do its job as it is only an 
implication. I don't think it can be written in the other direction (RIF 
expert needed here).
- three different vocabs are needed to represent three correspondences.

An alternative solution is to use EDOAL [1] which is made for alignment 
representation and allows export to skos, owl or rif through the 
Alignment API [2].
Here are the above three correspondences in EDOAL (rdf/n3):
:1 a align:Cell;
  align:entity1 wo:scientificName;
  align:entity2 biol:name;
align:relation "=".

:2 a align:Cell;
  align:entity1 gs:hasScientificName;
  align:entity2 wo:scientificName;
  align:relation "<".

:3 a align:Cell;
  align:entity1 gs:hasScientificName;
  align:entity2 [
   edoal:and wo:scientificName;
edoal:and [
    a eodal:PropertyDomainRestriction;
    edoal:class wo:Species.
   ]
  ]

Hope this helps !


François

[1] http://alignapi.gforge.inria.fr/edoal.html
[2] http://alignapi.gforge.inria.fr/

is that they are equivalent but only for the part where

Le 09/02/12 21:18, Peter DeVries a écrit :
>  I was wondering what the best way to align similar ontology terms are 
> not exactly equivalent either because they have a specific domain or
>  because there are slight differences in how they are used. For 
> instance some uses of Scientific Name include form with authority and 
> others
>  have a separate term for the scientific name with authority.
>  Here are the terms that I am most concerned about, but similar issues 
> apply to genus, family etc.
> <!-- http://lod.taxonconcept.org/ontology/txn.owl#scientificName -->
>
> <owl:DatatypeProperty 
> rdf:about="http://lod.taxonconcept.org/ontology/txn.owl#scientificName">
> <rdfs:label>scientificName</rdfs:label>
> <vs:term_status>testing</vs:term_status>
> <owl:equivalentProperty 
> rdf:resource="http://lod.taxonconcept.org/ontology/txn.owl#hasScientificName"/>
> <skos:closeMatch rdf:resource="http://purl.org/NET/biol/ns#name"/>
> <skos:closeMatch 
> rdf:resource="http://www.uniprot.org/core/scientificName"/>
> <skos:closeMatch 
> rdf:resource="http://rdf.geospecies.org/ont/geospecies#hasScientificName"/>
> <skos:closeMatch 
> rdf:resource="http://rs.tdwg.org/dwc/terms/scientificName"/>
> <skos:closeMatch 
> rdf:resource="http://purl.org/ontology/wo/scientificName"/>
> <rdfs:isDefinedBy 
> rdf:resource="http://lod.taxonconcept.org/ontology/txn.owl"/>
> <rdfs:subPropertyOf 
> rdf:resource="http://www.w3.org/2004/02/skos/core#prefLabel"/>
> </owl:DatatypeProperty>
>
> <!-- http://lod.taxonconcept.org/ontology/txn.owl#commonName -->
>
> <owl:DatatypeProperty 
> rdf:about="http://lod.taxonconcept.org/ontology/txn.owl#commonName">
> <rdfs:label>commonName</rdfs:label>
> <vs:term_status>testing</vs:term_status>
> <rdfs:isDefinedBy 
> rdf:resource="http://lod.taxonconcept.org/ontology/txn.owl"/>
> <skos:closeMatch rdf:resource="http://purl.org/NET/biol/ns#commonName"/>
> <skos:closeMatch 
> rdf:resource="http://rs.tdwg.org/dwc/terms/vernacularName"/>
> <skos:closeMatch 
> rdf:resource="http://rdf.geospecies.org/ont/geospecies#hasVernacularName"/>
> <skos:closeMatch rdf:resource="http://purl.org/ontology/wo/commonName"/>
> <rdfs:subPropertyOf 
> rdf:resource="http://www.w3.org/2004/02/skos/core#altLabel"/>
> </owl:DatatypeProperty>
>
> <!-- http://lod.taxonconcept.org/ontology/txn.owl#authority -->
>
> <owl:DatatypeProperty 
> rdf:about="http://lod.taxonconcept.org/ontology/txn.owl#authority">
> <rdfs:label>authority</rdfs:label>
> <rdfs:comment>The author and year of the species description 
> publication i.e (Baker, 1899)</rdfs:comment>
> <vs:term_status>testing</vs:term_status>
> <rdfs:isDefinedBy 
> rdf:resource="http://lod.taxonconcept.org/ontology/txn.owl"/>
> <skos:closeMatch 
>  rdf:resource="http://lod.taxonconcept.org/ontology/txn.owl#scientificNameAuthorship"/>
> <skos:closeMatch  rdf:resource="http://purl.org/NET/biol/ns#authority"/>
> <skos:closeMatch 
>  rdf:resource="http://rs.tdwg.org/dwc/terms/scientificNameAuthorship"/>
> <skos:closeMatch 
>  rdf:resource="http://rdf.geospecies.org/ont/geospecies#hasScientificNameAuthorship"/>
> </owl:DatatypeProperty>
>
> Respectfully,
>
> - Pete
>
> ------------------------------------------------------------------------------------
> Pete DeVries
> Department of Entomology
> University of Wisconsin - Madison
> 445 Russell Laboratories
> 1630 Linden Drive
> Madison, WI 53706
> Email: pdevries@wisc.edu <mailto:pdevries@wisc.edu>
> TaxonConcept <http://www.taxonconcept.org/> & GeoSpecies 
> <http://about.geospecies.org/> Knowledge Bases
> A Semantic Web, Linked Open Data <http://linkeddata.org/>  Project
> --------------------------------------------------------------------------------------
Received on Friday, 10 February 2012 13:37:36 UTC

This archive was generated by hypermail 2.3.1 : Sunday, 31 March 2013 14:24:37 UTC