Approximating Shortest Paths in Spatial Social Networks

We evaluate an algorithm that efficiently computes short paths in social networks by exploiting their spatial component. The main idea is very simple and builds upon Milgram´s seminal social experiment, where target individuals were found by having participants forward, or route, messages towards the target. Motivated by the somewhat surprising success of this experiment, Kleinberg introduced a model for spatial social networks, wherein a procedure called ´greedy routing´ can be used to find short, but not necessarily shortest paths between any two individuals. We extend Klein berg´s greedy routing procedure to explore k>;=1 links at each routing step. Experimental evaluations on social networks obtained from real-world mobile and landline phone communication data demonstrate that such adaptations can efficiently compute accurate estimates for shortest-path distances.