Hello Lynda,

since these open datasets have been devised to expose their semantics explicitly the data naturally refer to existing vocabularies (also referred to as terminologies, ontologies) where the semantics of classes and properties (concepts and relations) is formally encoded via ontology languages, such as OWL (backed by Description Logics).

The data representation is built upon a data model (RDF) relying on the notion of triple which roughly corresponds to a simple sentence.

Each component is identified in the form of a URI.

The triple are building blocks for (large) knowledge graphs.

The predicate is the name for the linking relationship you mentioned in the question. Links may involve resources within the same dataset but also resources located in different datasets (hence the LOD cloud).

You can derive/extract specific knowledge for your applications by querying the graph(s) using SPARQL.

You can even manipulate data and the underlying ontologies programmatically (e.g. via Jena or OWLApi) and derive the knowledge you aim at employing an automated reasoner.

For a detailed insight you may refer to this book:

http://linkeddatabook.com

Thank you very much Mr Fanizzi, I really am grateful, I see clearly now.

So the predicate is the name of the semantic link, Thank you.

I will try to manipulate a LOD Dataset and interrogate it.

Thank you very much.