Abstract:Networks of dynamic systems, including social networks, the World Wide Web, climate networks, and biological networks, can be highly clustered. Detecting clusters, or communities, in such dynamic networks is an emerging area of research; however, less work has been done in terms of detecting community-based anomalies. While there has been some previous work on detecting anomalies in graph-based data, none of these anomaly detection approaches have considered an important property of evolutionary networks-their… Show more
“…G t is compared to G t−1 ) and a scoring function maps the graph pair to a real number, creating a time series of scores. The scoring function can be based on community detection [8,11,12], graph distance [10,17], tensor decomposition [20,21], compression [19], or other graph features. A threshold is then applied to the time series of scores to identify events (isolated abnormalities), or change points (time points where significant changes occur in the graph and then persist).…”
Dynamic graphs are a powerful way to model an evolving set of objects and their ongoing interactions. A broad spectrum of systems, such as information, communication, and social, are naturally represented by dynamic graphs. Outlier (or anomaly) detection in dynamic graphs can provide unique insights into the relationships of objects and identify novel or emerging relationships. To date, outlier detection in dynamic graphs has been studied in the context of graph streams, focusing on the analysis and comparison of entire graph objects. However, the volume and velocity of data are necessitating a transition from outlier detection in the context of graph streams to outlier detection in the context of edge streams-where the stream consists of individual graph edges instead of entire graph objects.In this paper, we propose the first approach for outlier detection in edge streams. We first describe a high-level model for outlier detection based on global and local structural properties of a stream. We then propose a novel application of the Count-Min sketch for approximating these properties, and prove probabilistic error bounds on our edge outlier scoring functions. Our sketch-based implementation provides a scalable solution, having constant time updates and constant space requirements. Experiments on synthetic and real-world datasets demonstrate our method's scalability, effectiveness for discovering outliers, and the effects of approximation.
“…G t is compared to G t−1 ) and a scoring function maps the graph pair to a real number, creating a time series of scores. The scoring function can be based on community detection [8,11,12], graph distance [10,17], tensor decomposition [20,21], compression [19], or other graph features. A threshold is then applied to the time series of scores to identify events (isolated abnormalities), or change points (time points where significant changes occur in the graph and then persist).…”
Dynamic graphs are a powerful way to model an evolving set of objects and their ongoing interactions. A broad spectrum of systems, such as information, communication, and social, are naturally represented by dynamic graphs. Outlier (or anomaly) detection in dynamic graphs can provide unique insights into the relationships of objects and identify novel or emerging relationships. To date, outlier detection in dynamic graphs has been studied in the context of graph streams, focusing on the analysis and comparison of entire graph objects. However, the volume and velocity of data are necessitating a transition from outlier detection in the context of graph streams to outlier detection in the context of edge streams-where the stream consists of individual graph edges instead of entire graph objects.In this paper, we propose the first approach for outlier detection in edge streams. We first describe a high-level model for outlier detection based on global and local structural properties of a stream. We then propose a novel application of the Count-Min sketch for approximating these properties, and prove probabilistic error bounds on our edge outlier scoring functions. Our sketch-based implementation provides a scalable solution, having constant time updates and constant space requirements. Experiments on synthetic and real-world datasets demonstrate our method's scalability, effectiveness for discovering outliers, and the effects of approximation.
“…The temporal dynamics plays a vital role while integrating provides a better perceptive of network behavior [1][2][3]. Basically, the community relates the grouping of nodes with a cluster connected with many edges and cluster exists with few edges [4].…”
Section: Introductionmentioning
confidence: 99%
“…The networks are becoming wider and wider since it is the period of information explosion. Thus, we required many effective community detection algorithms for analyzing the networks with millions of vertices [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15].…”
Abstract:The complex networks are offering a high resource of heterogeneous data and the proper and efficient analysis discovers the unknown information and relations in networks. Due to the huge number of users and nonfamiliar fraud detection system in complex networks, a lot of online frauds introduce to affects the networks. In this paper, we concentrate on both community and fraud detection to minimize the link and node failures in the complex networks. A hybrid optimization algorithm proposed for community and fraud detection in the complex networks (HCFD-Net). The first contribution is to detect the community based on fruit fly optimization algorithm with differential evolution (FOADE). The second contribution is that the fraud detection is achieved by contingency table terminology with multi-link metrics. The performance of the HCFD-Net is analyzed on different five real-world networks are Zachary's karate club, Bottlenose dolphins', American college football, American political books, and Amazon online purchase network. The simulation result shows that the proposed HCFD-NET perform very efficient than existing algorithms in terms of normalized mutual information (NMI) and network lifetime.
“…A link between two nodes exists if there is a significant statistical interdependence between their time series. Typically, the linear cross-correlation function is used as the simplest measure of the statistical interdependence of temporal series [11].…”
Abstract:In this study, we present results of applying data mining techniques on meteorological dataset obtained from the Institute of Hydrometeorology and Seismology of Montenegro. The dataset covers the measurements taken from all 11 main meteorological stations in Montenegro for the period 2010-2015. We build new climate classification system based on decision tree. The system is simpler (i.e. uses fewer attributes) and more accurate than the well-known Köppen climate classification system. In addition, we propose a novel procedure for climate network construction. Finally, we identify the regions within the same climate type in Montenegro's climate network with the Girvan-Newman algorithm for community detection and achieve better results with respect to classical K-means and hierarchical clustering algorithms.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.