Cognitive Studies in the Interpretation of Social Media Data: TextAnalyst and ChatGPT

This article is an expanded version of the report given at the National Conference on Artificial Intelligence with international participation (Smolensk, October 16–20, 2023). The work presents some results of a comparative study of the advantages and limitations of expert analysis using the TextAnalyst 2.3 neural network technology and analysis carried out using ChatGPT Plus on social media data generated by actors from different backgrounds to identify and interpret audience reactions. Data collection was carried out from October 1, 2022 to March 31, 2023 in social media VKontakte, Telegram, YouTube, and Dzen. The dataset contains 22 199 085 tokens. Data analysis showed that it is most effective to use a combined approach, taking into account the results of both these approaches.