Using online update of distributional semantics models for decision-making support for concepts extraction in the domain ontology learning task
Антон Викторович Аникин, Alexander Katyshev, Mikhail Denisov, V Smirnov, Dmitry Litovkin · IOP Conference Series Materials Science and Engineering · 2019
Most of the information processed by computer systems is presented in the form of text corpuses. The number of such texts (as well as the corpus as a whole) only increases with time, and therefore the word processing tasks remain relevant to this day. Ontology allows to describe semantics using domain concepts and relations between them [1, 2]. In the ontology learning task, the ontology is dependent on quality of corpus which may not be readily available. There are different approaches to creating ontologies (including the use of different tools and frameworks). This paper discusses the use of word2ve c (group of related models that are used to produce word embeddings) using online vocabulary update and extension of the original data corpus with additional training for the domain concepts extraction to automate the domain ontology creation.