Hi Peter,

Thank you for your detailed report.

The DBpedia ontology is (a) crowdsourced and (b) follows a data-driven 
approach. Classes and properties are mainly derived from the actual data 
coming from different Wikipedia chapters.
Those are the main reasons of the issues you mentioned.

It would be great if you could contribute a deep analysis and detect the 
inconsistencies.
In this way, we could clean the ontology up and provide rock solid 
semantics.
As you already mention lots of examples, a brand new ontology and exact 
deltas with the current one would be highly beneficial.

Cheers!

On 4/10/14, 12:17 AM, Patel-Schneider, Peter wrote:
> I did a quick analysis of the classes in the DBpedia ontology and found
> quite a few issues that I think need attention.
>
> -  Many classes have no instances.  Each of these empty classes
>     should be examined to see whether they should be removed or modified.
>
> -  The sports-related groupings are differentially populated, differentially
>     organized, and unaxiomatized.  These groupings should be regularized and
>     minimal axiomatizations provided for them.  For example, there would be
>     classes for Basketball under at least SportsLeague, SportsTeam, Coach,
>     and SportsEvent each defined as the restriction of the grouping elements
>     related to Basketball.  The sports groupings include Sport (which is
>     special), SportsLeague, SportsTeam, Athlete, Coach, SportsTeamMember,
>     SportsManager, SportsEvent, SportFacility, SportCompetitionResult,
>     SportsSeason, and Tournament.
>
> -  Numerous stated inclusion relationships are not correct when considering
>     the normal definition of the class names.  Each of these should be
>     examined and either descriptions of the classes that support the
>     inclusion relationship be provided or the relationship itself modified.
>     For example, instances of the RecordOffice class do not appear to be
>     non-profit organizations.  Some other examples of questionable or
>     outright incorrect subclasses here are TermOfOffice, BackScene,
>     ChessPlayer, PokerPlayer, TeamMember, Saint, FictionalCharacter,
>     MythologicalFigure, OrganisationMember, Religious, Baronet, Medician,
>     Professor, Embryology, Lymph, Constellation, Galaxy, ElectionDiagram,
>     Olympics, OlympicEvent, ControlledDesignationOfOriginWine,
>     PublicServiceInput, PublicServiceOutput, and ProgrammingLanguage.
>
> -  Some class relationships are missing.  For example, TeamMember is
>     unrelated to SportsTeamMember even though they are both supposed to be
>     members of athletic teams.  Some other examples of missing relationships
>     are between BullFighter and Bullfighter, between Host and TelevisionHost,
>     and between Comic and Comics.  The missing relationships should be
>     provided or the classes merged.
>
> -  Place is a rather unnatural union.  It should either be removed or
>     better organized.
>
> -  There are quite a few subclasses of Building that are not truely
>     buildings, including AmusementParkAttraction, Casino, Factory, Hotel,
>     MilitaryStructure, Abbey and the other religious places of worship,
>     Restaurant, ShoppingMall, and Venue.  Similarly, there are a number of
>     subclasses of ArchitecturalStructure that may not be architectural
>     structures, including Garden, PublicTransitSystem, and Park.  There are a
>     few subclasses of NaturalPlace that are not necessarily natural
>     places, including Canal, and even Lake.   These classes should be
>     moved up in the ontology.
>
> -  The subclasses of Species are not collections of species.  The
>     subclasses should either be modified or moved elsewhere in the
>     ontology.
>
> -  The normal definition of PopulatedPlace is much too narrow to encompass
>     all its subclasses.  A new general class should be created to
>     encompass the subclasses and PopulatedPlace be modified as necessary.
>
> - There are a number of strange top-level or second-level classes.  These
>     classes should be examined to ensure that they make sense.  Many of
>     these classes appear to be somehow related to measurements, including
>     Altitude, Area, Blazon, ChartsPlacement, Demographics, Depth,
>     GrossDomesticProduct, GrossDomesticProductPerCapita,
>     HumanDevelopmentIndex, Population, Sales, Statistics, and Tax.  Other
>     strange classes include LifeCycleEvent, Imdb, Listen, PenaltyShootOut,
>     PersonFunction, PoliticalFunction, Profession, TopicalConcept, Type, and
>     YearInSpaceflight.
>
> Even if I had editing rights to the ontology I think that the fixes I have
> outlined above go beyond what should be done without some discussion.
>
> Comments?
>
> peter
>
>
> ------------------------------------------------------------------------------
> Put Bad Developers to Shame
> Dominate Development with Jenkins Continuous Integration
> Continuously Automate Build, Test & Deployment
> Start a new project now. Try Jenkins in the cloud.
> http://p.sf.net/sfu/13600_Cloudbees
> _______________________________________________
> Dbpedia-discussion mailing list
> Dbpedia-discussion@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
>

-- 
Marco Fossati
http://about.me/marco.fossati
Twitter: @hjfocs
Skype: hell_j

------------------------------------------------------------------------------
Put Bad Developers to Shame
Dominate Development with Jenkins Continuous Integration
Continuously Automate Build, Test & Deployment 
Start a new project now. Try Jenkins in the cloud.
http://p.sf.net/sfu/13600_Cloudbees
_______________________________________________
Dbpedia-discussion mailing list
Dbpedia-discussion@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion

Reply via email to