An Asymmetric Hidden Metric Model for Complex Networks

Models of complex networks are important tools for understanding many real life systems that are best investigated as a set of connected entities. The hidden metric model is one such model. It assumes that the network topology is shaped by a hidden metric space where the nodes reside. This paper extends this model to the case where the distance perceived by one node towards another is different from that perceived in the opposite direction. This situation is encountered in many real networks such as social networks and asymmetric communication networks. Simulation affirms that although the absence of symmetry does not affect the network, the loss of transitivity deteriorates clustering. Hence, in the proposed model, the metric space is replaced by an asymmetric metric space. The experimental results show that the proposed model generates networks with clustering coefficients comparable to those generated using the original model even though symmetry is lost.