Property based Schema Discovery for RDF using density based clustering[go to overview]
A problem in the Linker Open Data domain is that of incomplete schema specifications. Data sources do not all publish or refer to schemas capturing the whole of their data completely. Many approaches have been proposed for alleviating this issue. Recently, Kellou-Menouer & Kedad  discussed such an approach building on the property based clustering of entities into appropriate classes. They also show the discovery of class hierarchies and the domain-/range-typing for links. This talk will showcase their overall idea followed, point to future work and close in discussing the applicability for scenarios as with, e.g., the Koldfish project.
 Kellou-Menouer & Kedad (ER 2015): "Schema discovery in RDF Data Sources"
07.04.16 - 10:15