In our days, the Semantic Web has gained a lot of popularity since it provides standard ways regarding sharing and retrieving data. A key component of the Semantic Web is Ontology Engineering where it includes the tasks of ontology development and ontology alignment. These kinds of tasks require extensive human labor and a profound domain knowledge. There is a great need of automated solutions in Ontology Engineering. Ma chine Learning techniques are applied to various domains in order to provide experts with such solutions. This thesis investigates the application of Machine Learning in Ontology Engineering by applying such techniques in the domain of e-Government and particularly on European Union’s Vocabularies. Specifically, this thesis has two objectives: a) Solve the problem of “Sub-property Link Prediction” in an ontology set. b) Introduce an “Ontology Search Tool” based on pre-trained vector representations (text-embeddings). The experimental results are inspiring and indicate that Machine Learning techniques are applicable in On tology Engineering.
Collections
Show Collections