Cross-Platform Emotion Analysis and Hot Word Mining Method Based on Live Broadcast Room Bullet Screen

With the vigorous growth of the live streaming industry, the live streaming barrage data encompasses a considerable amount of valuable user emotion and topic information. This paper centers on the three live streaming platforms of Douyin, Kuaishou, and Xiaohongshu. By applying the methods of emotion analysis and hot word extraction, it conducts a comparison and analysis of the emotional tendencies and hotspots of users from different platforms in live streaming interactions. The findings reveal that users of Douyin have more positive emotions and focus on entertainment stars and fashion topics. Users of the Kuaishou platform have neutral emotions and are concerned about life services and local culture; while the emotions of users are more negative, and they are interested in skin care, beauty, and lifestyle. The cross-platform analysis approach proposed in this paper can offer data support for the optimization of live streaming content and user profiling.