Information Cascades Blocking Through Influential Nodes Identification on Social Networks.

Li,Xiaohua Zheng,Jing Han,Fei Hao

DOI: https://doi.org/10.1007/s12652-022-04456-x

IF: 3.662

2022-01-01

Journal of Ambient Intelligence and Humanized Computing

Abstract:Information cascades are identified as a major factor in almost every disastrous social network phenomenon, ranging from viral marketing, rumor spread, cyber violence, and various negative information diffusion. A small fraction of nodes has an important impact on the spread of information, therefore, mining these influential nodes helps block information cascades. Most of the existing influential nodes identification relies on topological features, however, they fail to simultaneously capture network topology, user behaviors, and information content during the diffusion, which may result in unsatisfactory control performance. To this end, we develop an effective method for blocking information cascades, which is based on extended bridge nodes identification. Individual decision-making behaviors with similar tendencies contribute to information cascades, but tight clusters block the diffusion of cascades. Bridges with double-edged characters on social networks are powerful ways to propagate new information from different clusters, but they are weak at forwarding information that is in some way risky or costly to adopt forwarding behaviors need to be supported by a higher proportion of neighbors. Threshold models for diffusion and double-edged bridges have been used to motivate the definition of bridge_like nodes and the modeling of information cascades in this article. A blocking method is proposed to identify as few influential extended bridge nodes as possible in a network to maximize information cascades blocking, and finally, the performance of the blocking method is validated on social networks.