In this paper, we have presented a structured review of the
various methods proposed for anomaly detection in social net- works represented as graphs. Mining
social networks for anoma- lies is a challenging and computationally intensive task due to the huge
size of the network and its dynamic nature. In the past decade, there are a wide variety of methods
developed for social network anomaly detection in different problem settings. This paper organizes
the state-of-the-art methods into different cate- gories based on the elementary approach followed
by each method and briefly introduces the corresponding methods. Finally, we have discussed the
various research challenges and open issues for future research in this domain. With the different
methods discussed, deciding a particular algorithm for anomaly detection is a difficult task. When
selecting an appropriate algorithm, one has to consider different aspects of the application such
as the type of the network being examined and the types of anomalies to be detected. This
comprehensive review provides a better under- standing of the various techniques that have been
developed for
mining social networks for anomalies.