Revealing the co-occurrence patterns of public emotions from social media data
This study proposes a framework that integrates deep learning with co-occurrence theory to classify public emotions from social media data into single, dominant-subordinate, and compound classes. The method identifies 34 types of complex emotions and reveals associations between urban environmental factors and emotional distribution. Compared to valence-based methods, it effectively captures changes in public emotions following a respiratory infectious disease outbreak, demonstrating robust stability across spatial scales.
Mar 11, 2026
