| Abstract |
I present the results of our study on how the decentralized nature of the LOD, and the explicit and mixed ontological model of the RDF data can affect the inference process. To do that, we analyzed the inference results of more than 600K individual RDF data documents. The results show that the inference process on the majority of data documents do not yield much, particularly because most of the ontological information that is needed to infer new knowledge is missing. We also analyzed the two mechanisms that the Semantic Web offers to tackle this issue, namely the owl:imports, and the dereferencability of IRIs. In the talk I will also present how much the inclusion of owl:imports ontologies, and IRI dereferencability can help to improve the situation. |