Problem Domain Ontology Mining Based on Distributional Semantics
Elena B. Kozerenko, Yulia Sinyaghina, Н. В. Сомин, Svetlana Pozdeeva · 2019
The paper presents the method of creating an extended ontology as an associative portrait of a subject area and the construction of a semantic space for intelligent knowledge extracting systems development. The ideology of semantic contextual spaces is based on the Distributional hypothesis: linguistic items with similar distributions have similar meanings. The model employed uses an extended hypothesis which includes the study of similarities and differences in contexts not only for a separate lexeme, but also for arbitrary multi-word text fragments which are considered as meaningful phrases.