The Impact of the Collective Influence of Search Engines on Social Networks

A social network contains a significant set of spreaders whose activities can lead to largescale activation of network members. In order to find the minimal set of spreaders, many methods based on traditional network topology have been proposed. However, search engines change the structure of traditional social networks. With the help of a search engine, each spreader has the potential to establish connections with disconnected spreaders. Thus, it is necessary to take the influences of search engines into account, in order to find a more accurate set of spreaders. In this paper, we aim to quantitatively characterize the impact of the collective influence of a search engine on a dynamic social network. First, we design a model to specially describe connections established by a search engine. Second, we improve a method based on collective influence theory to identify a more optimal set of super-spreaders, taking the influence of the search engine into consideration. We use the number of probably established subcritical paths attached to a node as this node's contribution in this social network. Third, we propose an algorithm based on collective influence that is applicable to networks with search engines to identify the optimal set of spreaders. The analysis results from both randomly generated networks and real-world networks indicate that our method can yield a more accurate set, which can cause a more large-scale cascade of information.