Community detection is an important topic in complex networks. A bipartite network is a special type of network, whose nodes can be divided into two disjoint sets and each edge connects between different types of nodes. In bipartite networks, there are two types of community definition, one-to-one correspondence and many-to-many correspondence between communities. The latter is better to represent realistic community structure in bipartite networks. However, few method can extract this type of structure. In this paper, we propose BiMLPA, based on multi label propagation algorithm, to detect many-to-many correspondence between communities in bipartite networks. Experimental results on real-world networks show that BiMLPA is effective and stable for detecting communities.
各種検索
サポート
ヘルプ