Centrality-Burst Detection in Social Networks: An Efficient Approach for Data Stream

In large social networks, being able to identify the key members, or so called central members, is one of the most important issues. Such members could be a good starting point for further analyzing. For example, the key members' activities with regard to the targeted products could be expanded to help marketing, or personalization advertising could be targeted to them with priority. However, with a "big velocity" and the complexity of the graph-structure of the data in social networks, identifying of the central members must be performed with an appropriate and efficient approach. In this paper, we propose an approach to identify the centrality of the social networks using the concept of burst detection in the streaming data environment. First, we present the definition of the centrality-burst in the problem setting. Then, an efficient streaming algorithm with QUBE technique is proposed. The efficiency of our work is also evaluated by experiment results. It is found that the proposed work is highly efficient. In addition, a simple approach to adjust parameters for the proposed approach is illustrated.