Sentimental reflection of global crises: Czech and Ukrainian views on popular events through the prism of internet commentary

Abstract

Social media have become a part of our lives, and their use helps us learn about events and comment on them with certain emotions. The purpose of our study was to determine the most frequent tone (positive, negative, neutral) of comments on impactful emergency and crisis news in the Czech Republic and Ukraine on a specific topic (pandemics, war, natural disaster etc.) using the sentiment analysis method. The methods of the study included a theoretical analysis of literature, social media (Twitter, Telegram), a Python program using: large language models GPT-3.5-Turbo and Twitter-XLM-RoBERTa, processing and interpretation of results (psycholinguistic).