sones GraphDB identifies correlations between Wikipedia entries

Leipzig, 4 April 2011 - The new “sones GraphDB” database analyzes Wikipedia entries in order to optimize connections between these entries. To achieve this research goal, sones GmbH is making its application available to the University of Leipzig. The novel database will analyze some 3.5 million data records for the DBpedia project.

Improved linking of Wikipedia articles

DBpedia is a joint project by the University of Leipzig, the Free University of Berlin, and Open Link Software. It extracts structured knowledge from Wikipedia articles and combines it with other information, such as the New York Times news site and the Project Gutenberg book portal. The new method will help improve consolidation of existing information. In the past, DBpedia primarily used semantic technologies for this. Now it will be possible to find even indirect and remote connections between entries. “The previously used technologies are efficient when it comes to administering and querying large data volumes,” says Sören Auer, Project Manager at the University of Leipzig. “However, it is very difficult to find connections between the entries across multiple information nodes. sones GraphDB is perfect for this task, and we are currently evaluating its performance as the first of its kind to take on this challenge.”

Users detect connections faster

The goal of the cooperation is to better depict correlations between the respective Wikipedia articles. “Social networks already allow you to describe and analyze relationships between various people. In the future, it should be easier for Internet users to spot even indirect connections between travel offers and tourist destinations, for example, or between companies and politicians,” the Project Manager explains.

The future of the pharmaceutical and automobile industry

“GraphDB examines data records in which the connections were previously created solely manually by the authors,” says Daniel Kirstenpfad, CTO of sones. “The application also finds previously hidden correlations between the individual Wikipedia articles which can then be set using new links. In the future, it will thus be possible to better link information from other branches as well, such as the pharmaceutical or automobile industries.”

DBpedia structures knowledge

DBpedia analyzes Wikipedia articles based, among other things, on people, places, organizations as well as historical events and biological classifications. The result is a knowledge base where you can query linked data similarly to a database. This will allow questions to be answered for which individual Wikipedia articles have no response. In DBpedia, for example, you can search for all female mayors of cities with populations over 50,000 or for all soccer players with jersey number 11.

About GraphDB

GraphDB by sones saves and analyzes structured, semi-structured and unstructured data. To this end, it uses its proprietary Graph Query Language - GQL. It is based on the well-known SQL and was expanded by sones for practical applications based on graph theory.

About sones

sones GmbH is headquartered in Leipzig and was founded in Erfurt in 2007. In 2009, the company was fortunate enough to acquire T-Venture, Deutsche Telekom’s venture capital company, as an investor. sones GmbH specializes in developing graph databases, which are supported by the company’s own file system. The sones GraphDB database technology works similarly to the human brain: It combines information by linking complex, semi-structured data volumes. This creates scalability and increases performance when saving and analyzing data. The sones database technology is laying the foundation for many new applications, such as in the Internet, document management, and information lifecycle management areas. In 2011, sones GmbH was able to successfully complete another round of financing. Investing in the company were Technologiegründerfonds Sachsen (TGFS) as the main investor as well as T-Venture and the Kreditanstalt für Wiederaufbau (KfW) (Reconstruction Credit Institute). For more information and to download the open source edition, go to www.sones.com.