LInk discovery framework for MEtric Spaces @ IKS Paris Workshop

LIMES Logo

The Research Group Agile Knowledge Engineering and Semantic Web (AKSW) is hosted by the Chair of  Business Information Systems (BIS) of the  Institute of Computer Science (IfI) /  University of Leipzig as well as the  Institute for Applied Informatics (InfAI). It consists of the three subgroups Emergent Semantics, Machine Learning and Ontology Engineering, and Semantic Abstraction.

The Semantic Abstraction (SIMBA) Group’s focus lies on knowledge extraction, integration, linking and consumption for porting the document-oriented web to the Data Web. For this purpose, SIMBA develops novel and scalable approaches for RDF/OWL extraction, link discovery and graph analysis. In addition, SIMBA provides tools and frameworks that implement these approaches and allow for their swift integration into industry projects.

One of SIMBA’s core projects is LIMES, the Link discovery Framework for metric spaces.
The basic observation behind LIMES is that links between knowledge bases play a key role in important tasks such as answering complex business questions, large-scale inferences and semantic data integration. Given the size of the Web of Data and the rate at it grows, time-efficient Link Discovery frameworks are central to allow to address these tasks.

LIMES is (to the best of our knowledge) the fastest lossless Link Discovery Framework for the Linked Data Web. LIMES provides a modular architecture that incorporates functionality for:

  • processing data from several types of input sources (SPARQL endpoints, CSV and N3 files, etc.)
  • preprocessing and cleaning of  the input data (tranformation via regular expressions, reduction to numbers, etc.)
  • time-efficient link discovery in metric spaces and
  • serializing the output in several output formats (N3, TAB, etc.).

LIMES Architecture

All features are easily extensible to suit the user’s needs. Our approaches utilize the mathematical characteristics of metric spaces to filter out a large amount of those instance pairs that do not suffice mapping conditions specified by the user. By these means, LIMES can reduce the number of comparisons needed during the mapping process by several orders of magnitude.

During our presentation at IKS Paris Workshop on the second day July , we will focus on

  • the importance of Link Discovery,
  • the LIMES framework itself and demonstrate
  • several experiments conducted with the framework.

Further links
Demo: http://limes.aksw.org
More Information: http://limes.sf.net

Comments are closed.