Jernej Stegnar, Jože M. Rožanec, Gregor Leban and Dunja Mladenić

ABSTRACT
When building a causal graph from textual sources, such as media
reports, a key task is to provide an accurate semantic understand-
ing of the graph nodes and to link them to existing ontologies
with at least two purposes: (i) expand the knowledge with already
created ontologies and (ii) guarantee accurate and different levels
of abstraction of the extracted concepts. This article describes
how we used OntoGPT, a tool for matching raw text to ontology
concepts initially designed for the medical domain, to match con-
cepts from media events to relevant ontologies. In particular, we
developed a set of scripts to generate custom YAML templates
and Python code to facilitate the extraction of relevant concepts
from ontologies and link them to causality graphs. The article
discusses the operation of OntoGPT and the template generation
process and addresses the tool’s limitations encountered during
the abovementioned process. Given the interest in developing
a foresight tool that addresses strategic foresight needs in the
green energy domain, three ontologies related to energy and the
oil industry were considered.