Abstract

The generation of natural language from Resource Description Framework (RDF)data has recently gained significant attention due to the continuous growth ofLinked Data. A number of these approaches generate natural language inlanguages other than English, however, no work has been proposed to generateBrazilian Portuguese texts out of RDF. We address this research gap bypresenting RDF2PT, an approach that verbalizes RDF data to Brazilian Portugueselanguage. We evaluated RDF2PT in an open questionnaire with 44 native speakersdivided into experts and non-experts. Our results suggest that RDF2PT is ableto generate text which is similar to that generated by humans and can hence beeasily understood.

 

Quick Read (beta)