RE: SKOS exactMatch Inclusion - a question concerning ISSUE75

Dear Hong Sun,

 

Skos reference discusses property chains including skos:exatMatch in http://www.w3.org/TR/2008/WD-skos-reference-20080125/#L5675 

 

There is no entailment specified between

- any pair of skos:broadMatch, skos:exactMatch, skos:relatedMatch and skos:narrowMatch

- skos:exactMatch and skos:broaderTransitive (or skos:narrowTransitive)

- the skos matching/mapping properties in general and skos:exactMatch in particular are very useful because they do not imply owl:sameAs semantics.

  so if 2 concepts each from a different concept scheme have an exactMatch, they still need not be the same thing.

 

The entailment will not work by formal inference.

 

The suggested inferred relation may be a relevant one to check and then to assert by some other intelligence. 

You could indicate such interesting relationships by other properties than skos properties.

 

Kind Regards,

 

Johan De Smedt 

Chief Technology Officer

 

mail:  <mailto:johan.de-smedt@tenforce.com> johan.de-smedt@tenforce.com

mobile: +32 477 475934

mail-TenForce

 

From: Hong Sun [mailto:hong.sun@agfa.com] 
Sent: Wednesday, 02 October, 2013 12:31
To: public-esw-thes@w3.org
Cc: Giovanni Mels; Jos De Roo
Subject: SKOS exactMatch Inclusion - a question concerning ISSUE75

 

Dear All, 

I am using SKOS for terminology mapping, as well as mapping validation. 
I have a question concerning exactMatch inclusions: 
Is it correct, or at least is it conventionally correct to have the entailment below? 

<A> skos:broaderTransitive <B> . 
<B> skos:exactMatch <C> . 
...entails: 
<A> skos:broaderTransitive <C>. 

It is similay to the question raised in ISSUE 75 ( <http://www.w3.org/2006/07/SWD/track/issues/75> http://www.w3.org/2006/07/SWD/track/issues/75), 
where decisions are made that 
<A> skos:broadMatch <B> . 
<B> skos:exactMatch <C> . 
does not entail: 
<A> skos:broadMatch <C> . 

I can understand this, because skos:broadMatch itself is not transitive, 
<A> skos:broadMatch <B> . 
<B> skos:broadMatch <C> . 
also does not entail: 
<A> skos:broadMatch <C> . 

But meanwhile, with inference, 
{ <A> skos:broadMatch <B> } => { <A> skos:broaderTransitive <B> }. 
{ <B> skos:broadMatch <C> } => { <B> skos:broaderTransitive <C> }. 
it can be entailed: 
<A> skos:broaderTransitive <C> . 

Therefore, with the definition of skos:exactMatch 
'The property skos:exactMatch is used to link two concepts, indicating a high degree of confidence that the concepts can be used 
interchangeably across a wide range of information retrieval applications. skos:exactMatch is a transitive property.' 
I am wondering when both properties are transitive, like below 
<A> skos:broaderTransitive <B> . 
<B> skos:exactMatch <C> . 
can we entail ? 
<A> skos:broaderTransitive <C> . 


In addition, the resolution of ISSUE 75 is made as below, 
RESOLUTION: Close ISSUE-75 by asserting that there are no property chain axioms as there is no evidence yet to support them 

A use case I encountered is that in terminology mapping, e.g. between concept scheme A and B 
suppose it exists: 
<A1> skos:broader <A2>. 
<B1> skos:exactMatch <A1>. 
it would be useful to entail the relation between <B1> and <A2>. 

How's your opinion on this? Thanks! 

Kind Regards,

Hong Sun | Agfa HealthCare
Researcher | HE/Advanced Clinical Applications Research
T  +32 3444 8108