A scalable incremental algorithm for computing the evolution of structural virality in social networks

The analysis of social networks is a concern nowadays. The importance they acquired over the last few years forces one to analyze suspicious behaviors and how false information spreads. Nevertheless, there are many parameters to consider when analyzing these networks. When the content of the messages is taken into account, semantic analysis can be considered the most important element. However, the relationships established between users and their interactions can reveal the different tendencies of the social network. To analyze several aspects of the structure of the conversations, it is required to know how the cascades spread. This can be represented by the use of virality. This paper reveals a scalable algorithm that drastically reduces the complexity of the calculation of the virality evolution. With the use of graph properties, the algorithm is a key component, able to relate time and size, to characterize the shape of the conversations in social networks in a simple way.