MENU

Datasets for resolving range violations in DBpedia

A range violation error is a problem when an object of a triple does not have a type required by the range of the triple's predicate. Based on our observation, 18.7% of DBpedia triples whose predicate is a mapping-based object property with a defined range are suffering from this kind of error. These errors could be probably due to

  • The predicate is incorrect and should be replaced
  • The range of the predicate is incorrect and should be refined
  • The object is incorrect and should be replaced
  • The object misses the correct (required) type

In this work, we focus only on erroneous triples that are due to incorrect objects. For example,

Subject Predicate Object Required Type
dbr:Sedo dbo:locationCountry dbr:Cologne dbo:Country
dbr:Safi_Airways dbo:targetAirport dbr:New_Delhi dbo:Airport
dbr:Stewart_Steven dbo:employer dbr:Daily_Mail dbo:Organisation

One way to fix these errors is replacing the incorrect objects with the correct objects which have the required types. For instance,

  • dbr:Cologne ⇒ dbr:Germany
  • dbr:New_Delhi ⇒ dbr:Indira_Gandhi_International_Airport
  • dbr:Daily_Mail ⇒ dbr:Daily_Mail_and_General_Trust

Datasets

The eight datasets we publish on this page were used to evaluate our work [1]. Each of them is a csv file containing one hundred erroneous triples of a particular object property p (in DBpedia) together with their correct objects ans which we manually find based on the information from corresponding Wikipedia pages. The eight properties and their ranges are

Property p Range (rp)
dbo:locationCountry dbo:Country
dbo:language dbo:Language
dbo:targetAirport dbo:Airport
dbo:routeEnd dbo:RouteOfTransportation
dbo:formerTeam dbo:SportsTeam
dbo:college dbo:EducationalInstitution
dbo:employer dbo:Organisation
dbo:birthPlace dbo:Place

File Format

s,p,o,answer

Download

Dataset8-1.0

Contact

Piyawat Lertvittayakumjorn: mail address
Ryutaro Ichise: mail address

Reference

  1. Piyawat Lertvittayakumjorn, Natthawut Kertkeidkachorn, Ryutaro Ichise. Resolving Range Violations in DBpedia. In Proceedings of 7th Joint International Semantic Technology Conference (JIST). 2017