社交网络时代舆情预警的挑战、模式及趋势研究

Public Opinion Early-Warning in Social Media Era: Challenges, Models and Trends

文章从新闻传播格局的新变化分析社交网络时代舆情预警面临的挑战,总结了当前敏感词库预测、传播特征值预测、关键意见领袖预测、多元子数据库预测四种预测模式,并探讨了舆情预警面临的底层爬虫技术、预警指标的多寡以及准确度过低等一系列问题,认为社交平台短文本的语义分析、文本内容分析与社会关系分析相融合、地理位置与社会关系关联耦合等是未来舆情预警的发展趋势。

Based on the new changes in the pattern of news dissemination, relay communication, emotional communication and relationship communication, this article analyzes the challenges the public opinion early warning system are facing. It summarizes four predicted models： the prediction of the sensitive word bank, the prediction of the characteristic value of communication, the prediction of the key opinion leader, the prediction of the KOL and the prediction of the multiple sub database. It also discusses the preview of the public opinion. It argues that the semantic analysis of short text on social platform, the integration of content analysis and social relationship analysis and combination of geographic location information with social information represent the new trend of public opinion early warning.