A Multithreaded Algorithm For Mining Maximal Cohesive Dense Modules From Interaction Networks With Gene Profiles
BCB(2016)
摘要
Several graph datasets exist which have additional attributes, representing properties of either nodes or edges in the graph. Recent research has focused on finding integrated or cohesive clusters where the clusters are not only densely connected but also have similarities in a subspace of their attributes. Cohesive clusters are more robust and accurately represent the cluster structure as they exhibit similarity in two domains (network structure and attributes).In this paper, we propose a multithreaded enumeration technique, for mining maximal cohesive and dense clusters from node-attributed graphs. For relaxed constraints the number of result clusters on a large graph can be very high and we need a way to find a representative set of clusters which can be most similar to all remaining clusters. We propose a novel technique to find representative clusters from the output space. Experiments on two real interaction networks show the effectiveness of the proposed algorithm.
更多查看译文
关键词
Biological networks,Dense module enumeration,K-medoids,Reverse search
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络