Abstract: An effective way to deal with massive datasets is to decompose an algorithm into a series of irrelevant tasks,and then to execute them in parallel by using open source softwares.Among overlapping community detection algorithms,the methods based on local expansion in its expansion phase only need the information of local communities and their corresponding neighbors,thus they have the possibility to be executed in parallel.In this paper,we proposed a pa-rallelizable algorithm utilizing local expansion for overlapping community detection,and implemented it by using open source software Spark.The algorithm consists of four phases.Firstly,a group of irrelevant central vertices are selected and their corresponding local networks are used as seeds.Secondly,the algorithm filters the selected seeds by removing those whose vertices are weakly connected.Thirdly,the algorithm adopts a batch expansion strategy to expand seeds,by adding a group of neighboring vertices into the local community or removing a group of vertices from the local community.Finally,similar communities are merged.Experimental results based on artificial networks and real world networks show that our method is both accurate and efficient.

Key words: Complex network,Overlapping community detection,Local expansion,Parallelizable algorithm,Spark