چكيده به لاتين
Detection of communication outbreak among members of a network or a subgroup of a network have been a topic of interest for quite a long time in the research area of social network analysis (SNA). Fraud detection, act of terrorism detection, change detection in political networks, and other similar applications show the importance of outbreak detection inn social networks. One approach to monitor changes in a social network is to monitor graph measures related to the network representation in each time period and detecting anomalies by applying a control chart. In several papers, necessity of measures and control charts comparison have been discussed. In this paper, we compare the performance of average degree and standard deviation of degree measures of a network for detection of outbreaks on a weighted undirected network using exponentially weighted moving average (EWMA) and cumulative sum (CUSUM) control charts. Evaluation results indicate that average degree measure is better at detecting small changes than standard deviation of degree measure. Whereas, for greater changes and outbreaks consisting of more members of the network, the opposite is true. In addition, EWMA control charts perform better than CUSUM in detecting smaller changes and outbreaks consisting of less members of the network.
Keywords: Social network monitoring, Degree centrality measure, Exponentially weighted moving average (EWMA) control chart, Cumulative sum (CUSUM) control chart, Average run length (ARL)
