Anticipating Viral Topics on Twitter by Detecting Micro-Influencers, Using a Combination of Social Network Analysis and AI
Anticipating Viral Topics on Twitter by Detecting Micro-Influencers, Using a Combination of Social Network Analysis and AI

Summary/Abstract: Detecting potentially viral topics on Twitter has been the subject of numerous studies, and several monitoring platformsoffer alerts for emerging topics to their users. However, solutions based on semantic analysis of publications are often impreciseand ineffective. In this article, which results from a research project, we propose a methodology based on the application of AIto metrics from Social Network Analysis, which analyses the dynamics of exchanges on social networks. We identified ‘micro-influencers’ who are active on six societal topics. Micro-influencers are interested in new topics ahead of opinion leaders, and theiractivism allows them to be picked up outside their communities: they are therefore precursors to the virality of new emerging topicsin the public sphere. By applying AI to the dozens of metrics offered by the Gephi Social Network Analysis software, we defineda machine learning model capable of successfully identifying these micro-influencers. To do this, we used an innovative tool thatmakes it possible to compare the effectiveness of several dozen machine learning models

• Details
• Contents