Multidimensional social media sentiments and stock market volatility: a deep learning approach

Abstract This study examines the role of different social media sentiment dimensions in explaining stock market volatility in Pakistan. Using a comprehensive dataset of textual information extracted from Twitter, we construct multiple sentiment measures including sentiment intensity, disagreement, and attention. These measures are derived from sentiment analysis of tweets through a deep learning BERT framework and incorporated into conditional volatility models to assess their impact on market fluctuations. We found that sentiment intensity and disagreement positively and significantly drive stock market volatility. However, investor attention stabilizes the market instead of fueling volatility when modeled jointly with behavioral dimensions. We conclude that it is the sentiment and not mere volume of discussion that translates into market risk.