Extreme Value methods for networks

点击翻译按钮获取中文摘要

When a network grows, what is the extremal behaviour of the nodes? We would like to examine if there a correspondence between the distribution of the extremes of the node centrality measures, such as in-/out-degrees, and the growth rules of the networks. We are particularly interested in the preferential attachment model for networks (and variants thereof), as the degree extremes of real-life networks that the model seems suitable for are better fitted by extreme value methods. We will also extend the theory and methods to other kinds of networks.