Hi Rupert, Thanks for detailed explanation and pointers.
I'm looking forward to developing a FOAF profile based entity-linking and disambiguation module in Stanbol, possibly as a new EnhancementEngine. I think FOAF vocabulary is supported in Stanbol, but it's not used for 'Person' disambiguation as much (correct me if I'm wrong), This is also related to my interest in WebID (foaf+ssl) integration to support user management (WebID based authentication, ACL), and user network in Stanbol. My high-level Idea on FOAF based Entity (Person) Disambiguation process is as below, I'm still learning these concepts in depth so, I value your insights on this; 1. A data-input method to upload foaf profiles as Entities (entity generation using foaf profiles) and create a foaf ManagedSite. 2. Entity linking over foaf connections (knows, worksAt etc) 3 With large index of foaf based entities in entity-hub, execute 'Person' disambiguation algorithm ; - generate candidates pairs for disambiguation - co-reference algo on foaf entities (rule based or machine learning algo based on owl:sameAs etc) - identify co-referent foaf nodes and create clusters (iterative clustering of co-referent entitites) - Update ManageSite with co-referent entities 4. Entity Disambiguation Result Viewer Reference : Jennifer Sleeman, and Tim Finin. Computing FOAF Co-reference Relations with Rules and Machine Learning<http://ebiquity.umbc.edu/paper/html/id/506> ;* In Proc. **Proceedings of the Third International Workshop on Social Data on the Web * * * Thoughts are most welcome! Thanks, Dileepa On Fri, May 3, 2013 at 11:06 AM, Rupert Westenthaler < [email protected]> wrote: > The Stanbol Enhancer uses the Stanbol Entityhub for Entity linking and > the default configuration of the Stanbol Launcher include a small > index of DBpedia.org entities. > > However you can also > > * link you own Entities. e.g. by creating a ManagedSite [1] and using > the RESTful interface to upload your entities > * install one of the available indexes downloadable at [2] > * use the Entityhub Indexing tool to create an index of some other dataset > > The usage scenario "Working with Custom Vocabularies" [3] provides > more information on this. > > best > Rupert > > [1] > http://stanbol.apache.org/docs/trunk/components/entityhub/managedsite.html > [2] http://dev.iks-project.eu/downloads/stanbol-indices/ > [3] http://stanbol.apache.org/docs/trunk/customvocabulary.html > > On Thu, May 2, 2013 at 8:58 PM, MJ <[email protected]> wrote: > > The entity extractor uses dbpedia by default. > > > > On May 2, 2013, at 9:54 PM, Dileepa Jayakody <[email protected]> > wrote: > > > >> Thanks Rupert. > >> Sorry if my question was not clear, what I wanted to know was whether > >> Stanbol uses dbpedia or freebase or any other data-set to create the > >> entities. > >> > >> Thanks, > >> Dileepa > >> > >> > >> On Fri, May 3, 2013 at 12:17 AM, Rupert Westenthaler < > >> [email protected]> wrote: > >> > >>> Hi Dileepa > >>> > >>> I do not exactly know what you are referring to as "default knowledge > >>> base", but the stable and full launcher do include a Entityhub Site > >>> based on DBpedia.org (version 3.6) with ~43.000 Entities (those with > >>> the most page links in dbpedia). > >>> > >>> best > >>> Rupert > >>> > >>> On Thu, May 2, 2013 at 10:47 AM, Dileepa Jayakody > >>> <[email protected]> wrote: > >>>> Hi All, > >>>> > >>>> Can I please know what is the default knowledge base used by Stanbol ? > >>>> > >>>> Thanks, > >>>> Dileepa > >>> > >>> > >>> > >>> -- > >>> | Rupert Westenthaler [email protected] > >>> | Bodenlehenstraße 11 ++43-699-11108907 > >>> | A-5500 Bischofshofen > >>> > > > > > > -- > | Rupert Westenthaler [email protected] > | Bodenlehenstraße 11 ++43-699-11108907 > | A-5500 Bischofshofen >
