Study on the Causes and Platform-Specific Evolutionary Mechanisms of Negative Social Public Opinion

The spread of social public opinion has transcended temporal and spatial constraints, with negative public opinion particularly prone to spiraling out of control. To gain an in-depth understanding of its evolutionary patterns and underlying causes, this study focuses on analyzing the formation mechanisms and development pathways of social negative public opinion. The research framework comprises three phases: Employing fsQCA to identify causal configurations, selecting paradigmatic cases for cross-platform comment crawling, and establishing curated text corpora; Implementing fine-tuned BERT models for sentiment prediction to derive platform-specific affective trajectories; Conducting temporal thematic evolution analysis via BERTopic modeling, ultimately synthesizing negative sentiment propagation pathways. Synthesizing emotional dynamics, thematic shifts, and phase-dependent dissemination patterns, platform-specific mitigation strategies emerge: Zhihu: Bolster mainstream media credibility and sustain monitoring during dissipation phases; Douyin: Enact emotion-containment protocols at outbreak initiation and restrict virality via content moderation; Weibo: Implement real-time surveillance of trending topics leveraging its Hot Search algorithm.