基于贝叶斯网络的健壮社团检测

A Novel Algorithm for Detecting Stable Community of Complex Network Based on Bayesian Network

健壮社团是复杂网络社团结构中稳定部分,健壮社团发现是非常困难的;提出了一种基于贝叶斯网络推理的健壮社团发现算法,把健壮社团发现问题当做推理问题,构造一个贝叶斯网络,根据结点的度来设置贝叶斯网络相关参数,然后将某些内部联系特别紧密的网络结点设为证据结点,在贝叶斯网络中进行信度传播,得到在已知证据的情况下其余结点属于该健壮社团的概率,最后得到复杂网络中的所有健壮社团;对足球俱乐部网络(115个结点)和随机网络(128个结点)的测试结果表明所提方法能有效地检测出复杂网络中存在的健壮社团,具有较好的应用价值。

The stable community is the stable part of complex network community structure, it is difficult to detecte the stable community of complex network. A novel algorithm for detecting stable community of complex network based on bayesian network is proposed. We regard the problem of detecting stable community as reasoning problems, construct a bayesian, network, set bayesian network related parame- ters according to the node degrees, put some internal contact special close network node as evidence node, spread the credibility in the bayes- ian network, get the probability which a node belongs to a stable community, and obtain all stable communities of a complex network.. The test results of football club network （115 nodes） and the stochastic network （128 nodes） indicate that the proposed method can effectively detect stable communities existing in the complex network , and has a good application value.