Opinion data mining based on DNA method and ORA software

• Published in: Physica A 2018-01, Vol.490, p.1471-1480
• Main Authors: Tian, Ru-Ya, Wu, Lei, Liang, Xiao-He, Zhang, Xue-Fu
• Format: Article
• Language: English
• Subjects: Dynamic Network Analysis; Opinion data mining; ORA; Rumor intervention
• ISSN: 0378-4371; 1873-2119
• DOI: 10.1016/j.physa.2017.08.093

Public opinion, especially the online public opinion is a critical issue when it comes to mining its characteristics. Because it can be formed directly and intensely in a short time, and may lead to the outbreak of online group events, and the formation of online public opinion crisis. This may become the pushing hand of a public crisis event, or even have negative social impacts, which brings great challenges to the government management. Data from the mass media which reveal implicit, previously unknown, and potentially valuable information, can effectively help us to understand the evolution law of public opinion, and provide a useful reference for rumor intervention. Based on the Dynamic Network Analysis method, this paper uses ORA software to mine characteristics of public opinion information, opinion topics, and public opinion agents through a series of indicators, and quantitatively analyzed the relationships between them. The results show that through the analysis of the 8 indexes associating with opinion data mining, we can have a basic understanding of the public opinion characteristics of an opinion event, such as who is important in the opinion spreading process, the information grasping condition, and the opinion topics release situation. •Build a Dynamic Network Analysis model to mine online rumor data.•Use 8 ORA measures to analyze the relationships between public opinion components.•Give a case study to investigate the characteristics of public opinion.