Ontology Extraction from Compound Sentences in Hungarian Language
László Kovács, Erika Baksáné Varga, László Rostás · 2018
The implementation of knowledge extraction from text is a complex task involving many processing layers. For Hungarian language, the available semantic analyzers can provide only partial functionality. This paper focuses on the analysis of compound sentences in Hungarian language and a novel model is proposed to generate RDF description for the input sentences. The presented model generates a set of candidate interpretations, and a weighting method - that uses syntactic and semantic components - is applied to select the winner semantic graph. Based on the test experiences, the implemented framework provides an improved semantic model of the input compound sentences.