Evaluation of Media-Based Social Interactions: Linking Collective Actions to Media Types, Applications, and Devices in Social Networks

There is a growing number of opportunities for users to perform collective actions in social networks: Such collective actions engage users in correspondents social interactions. Although some models for representing users and their relationships in social networks have been proposed, to the best of our knowledge, these models do not explain what the underlying social interactions are. In previous work, we have proposed a human-readable technique for modeling and measuring social interactions, which resulted from users’ actions that involved, for instance, media types, interaction devices, and viral content. In our technique, social interactions are represented as behavioral contingencies in the form of if-then rules, which are then measured using an established data mining procedure. After being able to represent and measure a variety of social interactions, we identified the opportunity of transforming our technique into a method for capturing, representing, and measuring collective actions in social networks. In this chapter, we present our method and detail how it was applied to represent and measure social interactions among a group of 1,600 Facebook users over the period of 7 months. Our results report the link among actions (e.g., like), media objects (e.g., photo), application type (Web or mobile), and device type (e.g., Android).