Résumé
Cybercriminals are getting more intelligent with their tactics in cyberattacks; by using a fake social
media profile, they are capable of copying a legitimate profile and perform different scale attacks. In
social networks, there are other types of cyberattacks such as Compromised Profile, Malicious Links
and Content, Social Engineering, and Reconnaissance. Cascading impact in fact is not always caused
through sophisticated attacks as observed in the case of SolarWinds by accessing to customer data.
There are much simpler examples, one of which is the constant occurrence of business email compromise, or even cascading of cyberattacks in multilayer networks e.g., from social media into business operation. Multilayer networks are ones with multiple kinds of relations in multidimensional settings as an extension of the traditional networks. At the same time, we aim to explore the cascading impact of cyberattacks on multilayer social networks. This means how a cyberattack originated by using a fake social profile will cascade into all parallel multilayer networks. We use dynamic processes in multilayer networks to understand how the cyberattacks are propagated. We use ML algorithms to detect fake and nonfake profiles, e.g. via a dataset in twitter, then use SIR (susceptible, infected or removed) model as a base generating simulated cascades with the goal of comparing them with real ones to assess how realistic this model performs in multilayer networks, e.g. in a dataset including profiles in Google+ -Instagram – Twitter to see how fake profiles are cascaded into parallel social networks.