The following query obtains information about schools in http://opendatacommunities.org/sparql
PREFIX skos: <http://www.w3.org/2004/02/skos/core#>
SELECT DISTINCT ?id ?name WHERE {
?school a <http://statistics.data.gov.uk/def/geography/LocalEducationAuthority>
; skos:notation ?id
; skos:prefLabel ?name
}
Both SKOS properties are used in the target datasets at http://data.admin.ch/query/, but neither of them for the class http://schema.org/School, which would be the preferred one to have in the recommendation.
Currently, all the recommendations preserve both SKOS properties, causing the optimal query not to be recommended at all. Indeed, the optimal query has nothing in common with the original one except for the rdf:type property. That one would be:
PREFIX schema: <http://schema.org/>
SELECT DISTINCT ?id ?name WHERE {
?school a schema:School
; <http://gov.genealogy.net/ontology.owl#hasDenomination> ?id
; schema:legalName ?name
}
Find a way to lose shared properties (and detect when they should be lost) in the generalization, without causing a combinatorial explosion of the specialization tree (note: this is a bitch tough).
The following query obtains information about schools in
http://opendatacommunities.org/sparqlBoth SKOS properties are used in the target datasets at
http://data.admin.ch/query/, but neither of them for the classhttp://schema.org/School, which would be the preferred one to have in the recommendation.Currently, all the recommendations preserve both SKOS properties, causing the optimal query not to be recommended at all. Indeed, the optimal query has nothing in common with the original one except for the
rdf:typeproperty. That one would be:Find a way to lose shared properties (and detect when they should be lost) in the generalization, without causing a combinatorial explosion of the specialization tree (note: this is
a bitchtough).