Structural analysis of IWA social network

Abstract

Internet Water Army (IWA), a special group of online users, has more and more engaged our attention due to the negative effects caused by their irresponsible comments or articles. While most of related work focused on how to detect IWA using a classifier, there is a lack of analysis about the distribution and behavior characteristics of the special group. To address this issue, this paper constructs an IWA social network in which IWAs are core nodes, and preliminarily studied the network from its traits of structure and composition. Firstly, we crawled IWAs from a task posting website and extracted the relations between them and normal users from a social network site. Then, we applied two classical community detection algorithms FN and CPM on the IWA social network, and analyze the community detection result. Experimental results show that IWA social network is deserved to analyze and discuss, and we found some interesting phenomena which are very helpful to better understanding and monitoring the IWA accounts.