Similar Structures inside RDF-Graphs
Anas Alzogbi, Georg Lausen · 2013
RDF is the common data model to publish structured data on the Web. RDF data sets are given as subject-predicateobject triples and typically are represented as directed edgelabeled graphs. To make the information represented by such graphs comprehensible, RDF-schema (RDFS) provides concepts to define a class-structure as part of the given RDFgraph and thus supports a more abstract view on the data set. In this paper we follow a different approach and propose to make an RDF graph more comprehensible by reducing its size by partitioning to discover subgraphs which are similar with respect to their structure. The methods applied to derive a partition are based on bisimulation and agglomerative clustering. We demonstrate the usefulness of the approach by applying it on several synthetic and one real world RDF datasets.