Automatic Mapping of Upper-Level Ontology Classes (DOLCE) and Domain-Specific Ontology ITSMO
Andrey P. Khalov, Olga Muratovna Ataeva · 2025
This paper proposes a method for extending the top-level ontology DOLCE (DOLCE-lite version, referred to as TLO) to the domain of IT services without expert involvement. The main challenge addressed is the automatic mapping of classes in conditions of a small number of objects (<100) and the absence of annotated data. A review of existing approaches is conducted, their limitations are identified, and novel mapping methods are proposed, integrating embeddings and large language models. The suggested method achieved an 82.35% mapping accuracy when integrating DOLCE and ITSMO ontologies. As a result, the ITO-seed ontology was developed, containing linked classes from DOLCE and ITSMO, which can be utilized in further research and in building knowledge graphs for IT Service Management (ITSM) systems.