An overview and perspective on social network monitoring

Detection of anomalies in social networks has always been a topic of interest in the realm of social network analysis. However, in recent years, researchers from industrial process monitoring area have applied their methods for monitoring social networks. In this review paper, we discuss social network monitoring approaches that apply control charts to detect anomalies in social networks. We mention the characteristics of the approaches and will shed some light on different topics that can be considered for further investigation by interested researchers.

“…Some papers have reviewed methods of monitoring social networks . Savage et al described 4 kinds of anomalies: static unlabeled, static labeled, dynamic unlabeled, and dynamic labeled.…”

Section: Literature Reviewmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

An overview of dynamic anomaly detection in social networks via control charts

Noorossana¹,

Hosseini

Heydarzade

2018

“…As Woodall et al stated the application of social network monitoring methods is often illustrated using 2 types of networks such as terroristic ones like the “al Qaeda” network or that based on Enron e‐mail communications. Here, we select the real‐world Enron email data set to show the application of the proposed methods.…”

Section: Case Study: Enron Email Datamentioning

confidence: 99%

Phase I monitoring of social networks based on Poisson regression profiles

Fotuhi

Amiri

Maleki

2018

Nowadays, due to the increasing role of social networks in our daily life, monitoring and forecasting social trends have attracted the attention of many researchers. To the best of the authors' knowledge, the literature includes few studies of monitoring social networks. Existing researches have focused on analyzing only the existence of communications between people and have neglected to monitor the number of such communications. In this paper, first counts of communications between people are modeled using Poisson regression profiles. Then, 3 Phase I monitoring methods, extended T2, F, and a standardized likelihood ratio test method is suggested to detect step changes, drift, and outliers in the parameters of Poisson regression profiles. The proposed methods are evaluated via simulation studies in terms of signal probability criterion. The results show that in most out‐of‐control situations the standardized likelihood ratio test method outperforms the T2 and F methods. Then, a numerical example and a case study based on Enron email data are presented to illustrate the application of the extended methods.

“…Meanwhile, some studies have generalized these measures to enable them to be applied on weighted networks . In addition to measures, use of control charts has been one of the prominent methods to monitor social networks that leads to detection of anomalies . McCulloh and Carley discussed social network change detection (SNCD), which is a combination of SNA and statistical process control, by which small changes can be identified via monitoring network measures.…”

Section: Introductionmentioning

confidence: 99%

“…As stated by Savage et al, research is needed to determine which metric is suitable to satisfy the purpose of monitoring. Also, it is mentioned by Woodall et al that McCulloh and Carley and Azarnoush et al advocated use of CUSUM and EWMA charts but performance comparisons are needed. Furthermore, Ranshous et al divided network anomalies into 4 categories.…”

Section: Introductionmentioning

confidence: 99%

Performance evaluation of EWMA and CUSUM control charts to detect anomalies in social networks using average and standard deviation of degree measures

Hosseini

Noorossana

2018

Detection of communication outbreak among members of a network or a subgroup of a network has been a topic of interest in the literature of social network analysis. One approach to monitoring changes in a social network is to monitor graph measures related to the network representation in each period and detecting anomalies by applying a control chart. In this paper, we compare the performance of average degree and standard deviation of degree measures of a network for detecting outbreaks on a weighted undirected network using exponentially weighted moving average and cumulative sum control charts. Evaluation results indicate that average degree measure is better in 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, exponentially weighted moving average control charts perform better than cumulative sum in detecting smaller changes and outbreaks consisting of less members of the network.