From protein interactions to Facebook friends, a new algorithm for link prediction (Physical Review Research paper)

Can your Facebook friendships help in predicting your next phone call? While network science is a pivotal tool to characterize the structure of real world complex systems, link prediction algorithms can help to reconstruct networks from incomplete data sets and to forecast future interactions in evolving networks. Facebook, indeed, is a good example of how the second task may be routinely applied on online social networks.

In Link prediction in multiplex networks via triadic closure
, a new paper out in Physical Review Research, a team of ISI Foundation researchers proposes a novel link prediction algorithm, by generalizing the Adamic-Adar method (one of the most common and successful model for link prediction in social networks) to multiplex networks composed by an arbitrary number of layers, that encode diverse forms of interactions.

Scientists address some of the fundamental challenges and questions still unanswered in the link prediction field, like how can different kind of relational data be exploited to improve the prediction of new interactions, or how can link prediction algorithms optimize the information structured in a multiplex network representation. They show that the new metric outperforms the classical single-layered Adamic-Adar score and other state-of-the-art methods, across several social, biological and technological systems, paving the way for a deeper understanding of the role of different relational data in predicting new interactions. And thus providing a new algorithm for link prediction in multiplex networks that can be applied to a plethora of systems: from protein interactions to social networks.

Link prediction in multiplex networks via triadic closure”, Alberto Aleta, Marta Tuninetti, Daniela Paolotti, Yamir Moreno, and Michele Starnini, Physical Review Research, 16 November 2020,