复杂网络中社团个数的估计

Estimating the Number of Communities in Complex Network

以随机块模型为基础,并假设社团之间顶点连边的概率符合伯努利分布,通过积分的方法构造出社团个数的似然函数。根据复杂网络的一个观测值求出这个似然函数的最大值,从而得到社团个数的估计。并用人工随机网络和真实的网络进行了验证。

Based on stochastic block model,a likelihood function of the number of communities is proposed by integral,under the hypothesis that edge between two vertexes from different communities be Bernoulli distributed.With an observed network,an estimation of the number of communities is given by maximizing the likelihood function.The model is applied on both real and random networks.