Liberté, Égalité, Fraternité: a power study in signed networks
该研究提出一种基于条件关系的概率框架，通过最大熵方法计算网络中个体和群体的权力指数，并应用于一个中等规模的Kronecker网络。
Abstract Power in human societies is a central phenomenon. Even though, it took ages to understand it and – even more – to measure it. Only in the last decades attempts were made to model power relations and to assign respective power indices to actors in a network. The present work goes a step further. It measures power of actors and groups of actors in networks by means of conditional relations. In a probabilistic framework, such relations are specified as conditionals: Which actor receives power given that the adjacent actor has it, and which actor looses power given that the neighbour dominates. This pattern of power relations allows for an exact calculation of an actor’s and groups of actors’ power index. The new decision analytics tool for this is maximizing entropy for the whole net and evaluating each actor’s influence therein. The new concept is applied to a middle size Kronecker net of clans and subclans operating in a today’s society.