Amis (Slovenia)

Abstract Controversy encompasses content that draws diverse perspectives, along with positive and negative feedback on a specific event, resulting in the formation of distinct user communities. Research on controversy can be broadly categorized into two domains: controversy detection/quantification, which aims to measure controversy on a topic, and controversy explainability, which seeks to comprehend the reasons behind a topic’s controversial nature. This paper primarily contributes to the realm of controversy explainability. We conduct an analysis of topic discussions on Twitter from a community perspective, investigating the role of text in accurately classifying tweets into their respective communities. To achieve this, we introduce a SHAP-based pipeline designed to quantify the influence of impactful text features on the predictions of three tweet classi-fiers. Our approach involves leveraging various text features, including BERT , TF − IDF, and LIWC. The results, derived from both SHAP plots and statistical analyses, distinctly reveal the substantial impact of certain text features in tweet classification. Furthermore, our findings underscore the significance of this study and underscore the potential advantages of combining text and user interactions for a comprehensive understanding of controversy quantification.