Nomination Analysis of Media Profiles of Public Opinion Leaders in Various Cultural Environments Based on GPT

In this paper, we present a novel technique for analyzing the metaphorical and nominative profiles of public opinion leaders under conditions of intense mediatization. To achieve this, we employ GPT models to identify linguistic nuances and contextualize messages from social networks, thereby enabling their classification within the information field they generate. By utilizing the case study of Mikhail Gorbachev, our technique delves into how fundamental metaphors and cultural interpretative schemes shape the perception of these leaders. Additionally, we incorporate archetypal linguistic features across various geo-cultural contexts to enhance our understanding of the public’s rational and emotional attitudes towards Gorbachev. The study reveals that the proposed technique significantly improves comprehension of the current information landscape. It offers strategic insights for predicting and proactively responding to potential information trends. Moreover, our method highlights the polarization of opinions in different media platforms, illustrating how direct insults and commendations vary significantly between VKontakte and Telegram. This interdisciplinary approach leverages the capabilities of advanced language models while combining technical expertise with deep contextual knowledge. By fostering collaboration among specialists from diverse fields, we aim to create more nuanced and reliable analytical tools. Our findings suggest that continuous reassessment and refinement of the dataset are crucial for capturing dynamic shifts in public opinion and enhancing the accuracy of the bot’s algorithms. This paper contributes to media analysis by providing a robust framework for future research in public opinion profiling.