Investigation of Emotion Exchange Motifs in Bot/Human Interactions During Riot Events
In this paper, we analyze a data-set consisting of 4.5 million tweets related to four highly emotional events and investigate the interaction patterns that emerge when social bots communicate with human users. In particular, we propose an emotion-annotated n-layer multiplex network model to study the involvement of bots in the exchange of emotional messages. To this end, we study four events: a) the riots that happened during the 2017 G20 summit in Hamburg, b) the 2017 Charlottesville riot, c) the 2017 Catalonia independence referendum riot, and d) the riots that happened after the Philadelphia Eagles won the 2018 Superbowl.
In our analysis, we found that: 1) when identifying significant triadic patterns (motifs) in the the respective communication network, a number of star-like subgraphs emerge as representative and significant communication patterns, 2) bots tend to send messages of a higher emotional intensity, as compared to humans, 3) though bots predominantly send retweets, they also compose original messages that may potentially harm the reputation of the person they are targeted at, and 4) in contrast to previous findings, we found that during riot events up to 87.18 % of the involved bots actively engage in a direct communication with human users.
E. Kusen, M. Strembeck, Investigation of Emotion Exchange Motifs in Bot/Human Interactions During Riot Events, In: Proc. of the 5th International Conference on Social Networks Analysis, Management and Security (SNAMS), IEEE, Valencia, Spain, (2018)