作者: Elise Desmier , Marc Plantevit , Céline Robardet , Jean-François Boulicaut
DOI: 10.1007/978-3-642-33492-4_11
关键词:
摘要: We focus on the discovery of interesting patterns in dynamic attributed graphs. To this end, we define novel problem mining cohesive co-evolution patterns. Briefly speaking, are tri-sets vertices, timestamps, and signed attributes that describe local co-evolutions similar vertices at several timestamps according to set express trends. design first algorithm mine complete a graph. Some experiments performed both synthetic real-world datasets demonstrate our enables discover relevant feasible time.