Research Project

Support Wikipedia Authors with Editing Suggestions

One strength of Wikipedia, and a central factor for its growth, is that it does not restrict contributors. On the other hand, this leads to many inconsistencies within Wikipedia, especially between the 251 different language versions.

 

Extracting structured data from all 251 versions of DBpedia and interlinking this data with background knowledge from ontologies like OpenCyc, SUMO, or WordNet, allows different types of consistency checks. For instance:

 

Multi-Domain Ontology

DBpedia is one of the largest multi-domain ontologies that currently exist. Compared to other ontologies which usually only cover specific domains, are created by relatively small groups of knowledge engineers, and are very cost intensive to keep up-to-date as domains change, DBpedia has several advantages:

 

Document Classification, Annotation and Social Bookmarking

Terms from DBpedia can be used to annotate Web content. Compared to other subject hierarchies, like the classic ones used within libraries, DBpedia has the advantage that each subject is backed by a rich description including abstracts in 14 languages. Another advantage compared to static hierarchies is that DBpedia evolves as Wikipedia changes.

 

Mobile and Geographic Applications

DBpedia contains information about geographic locations and is interlinked with other geo-related data sources such as Geonames, the US Census, EuroStat, and the CIA world fact book. The data set contains geo-coordinates for many geographic locations which enable location-based SPARQL searches.

This makes DBpedia a valuable data source for location-based applications. DBpedia contains short abstracts about places which display nicely on mobile phones and PDAs.

 

Include DBpedia Data in Your Web Page

One nice thing about Wikipedia is that is kept up-to-date by a large community. Therefore, if you need a table on your Web page with, say, German cities, African musicians, Amiga computer games from the 90s, or whatever, you could generate this table with a SPARQL query against the DBpedia endpoint, and your table will stay up-to-date as Wikipedia changes.

Such tables can be implemented either by using JavaScript on the client, or with a scripting language like PHP on the server. The second option also allows you to cache query results.

 

Revolutionize Wikipedia Search

Wikipedia currently only supports keyword-based search and does not allow more expressive queries like "Give me all cities in New Jersey with more than 10,000 inhabitants" or "Give me all Italian musicians from the 18th century." This lowers the overall utility of Wikipedia.

One major application domain for the DBpedia data set is to enable sophisticated queries against Wikipedia, which could revolutionize the access to this valuable knowledge source.

Here are three prototypical search interfaces using different approaches to improve Wikipedia search:

 
Subscribe to RSS - Research Project