ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING

Corpus analysis of factors influencing the tone of the word "patriotism"

M. M. Sharnin

Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences

Abstract: The work presents an effective method for identifying positive and negative factors influencing the emotional tone of the word “patriotism” in a 10 GB corpus of texts from Telegram channel messages. Factors influencing tone are identified among semantic connections obtained using the Word2Vec neural network and tonal dictionaries. The method uses statistical patterns between words from tonal dictionaries that are part of associative groups of semantically similar words. The strongest positive factors in the corpus are morality, spirituality, love for the motherland, and humanism. The method allows to monitor the factors influencing the tone of the word “patriotism” in a large, dynamic corpus of texts, which can contribute to increasing the effectiveness of promoting patriotism as a traditional value in the information environment.

Keywords: corpus analysis, factors influencing tone, patriotism, semantic connections, neural network, tonal dictionaries.

Received: 06.11.2025
Revised: 17.12.2025
Accepted: 10.12.2025

DOI: 10.60797/IRJ.2025.162.15