Improved Epidemic Path Predictability in Complex Networks

We apply recent results on random walkers to the analysis of idealized epidemic outbreaks in scale-free networks. By replacing the node degree with the {\it random walk centrality} we observe a refined hierarchical cascade leading to a greatly enhanced predictability for the order of infected nodes. We confirm our model results on data from real-world Internet maps at the autonomous system level. The present results are highly relevant for the advancement of dynamic and adaptive strategies that aim to mitigate network attacks.