sub:assertion {
  sub:paragraph c4o:hasContent "To assess the domain coverage gain, we can exploit two signals: (a) the amount of produced novel data with respect to pre-existing T-Box properties and (b) the overlap with already extracted assertions, regardless of their origin (i.e., whether they stem from the raw infobox or the ontology-based extractors). Given the same Italian Wikipedia dump input dating 21 January 2015, we ran both the baseline and the supervised fact extraction, as well as the DBpedia extraction framework to produce an Italian DBpedia chapter release, thus enabling the coverage comparison." ;
    a doco:Paragraph .
}