The significant volume, as well as the semantic and lexicological diversity of Internet content, necessitates the creation of new methods for its computer analysis in order to identify potential threats and undesirable content, in particular, using artificial intelligence (AI) methods. This scientific problem can be solved with the use of neural network technologies, including frequency preprocessing of text arrays, justification of the structure and construction of a domain-oriented database of text data bodies, justification and experimental study of the architecture and macroparameters of a hybrid neural network based on the vector representation of pre-formed dictionaries of terms.

Experimental study of the functioning and quality of the developed artificial neural networks configurations according to the selected accuracy criteria

E Melikhova

Papers

Experimental study of the functioning and quality of the developed artificial neural networks configurations according to the selected accuracy criteria