Abstract:We study network anomaly detection problem with subgraph search and vertex classification techniques in Chung-Lu networks, where small anomalous networks are embedded in background networks. The general detectors traditionally utilize the spectral characteristics of whole networks to decide whether the anomalous networks exist or not. Moreover, many algorithms model the background networks with special random graphs, such as Erdős–Rényi random graphs. However, these methods may not achieve good detection performance because the spectral information of anomalous networks extracting from the relational data of the whole networks is subtle. Furthermore, the assumptions of the models limit the applications of the methods. In order to improve the detection capability of algorithms and relax the restriction of the background networks models, we develop two preprocessing approaches, referred to as subgraph search and vertex classification, and propose two detection algorithms, which can be applicable to the generalized network model (Chung-Lu random graph). By leveraging statistic features of priori data, the subgraph search method can obtain local network nodes that are more anomalous than the remainder. Moreover, the vertex classification is used to further distinguish the anomalous nodes from the local ones. And then, by using the relational data corresponding to the local nodes, the detection statistics are constructed to detect the anomalous network. Additionally, based on concentration of measure theory, the probability bounds of performance analysis are derived for the presented algorithms. Simulation examples are provided to illustrate that the proposed detectors can achieve better detection performance than the existing algorithms.

A Spectral Approach to Detecting Subtle Anomalies in Graphs

Subtle Anomaly Detection in Dynamic Networks Using Graph Spectra

Community-based anomaly detection using spectral graph filtering

A Detection Algorithm to Anomaly Network Traffic Based on Spectral Kurtosis Analysis

Anomaly Detection with Subgraph Search and Vertex Classification Preprocessing in Chung-Lu Random Networks

Near-optimal Anomaly Detection in Graphs using Lovasz Extended Scan Statistic

SubAnom: Efficient Subgraph Anomaly Detection Framework over Dynamic Graphs

A Framework for Spotting Anomaly

On Spectral Analysis of Directed Signed Graphs

Extreme Value Modelling of Feature Residuals for Anomaly Detection in Dynamic Graphs

Spectral Analysis Of K-Balanced Signed Graphs

A Spectral Clustering-Based Framework for Detecting Community Structures in Complex Networks.

An Efficient Framework for Detecting Evolving Anomalous Subgraphs in Dynamic Networks.

Spectral Algorithm for Community Detection under Heterogeneous Networks

Multi-representations Space Separation based Graph-level Anomaly-aware Detection

Anomaly Detection in Graph Signals with Complex Wavelet Packet Correlation Mining

Community-based Outlier Detection for Edge-attributed Graphs

An Embedding Approach to Anomaly Detection

Exploiting Anomalous Structural Nodes in Dynamic Social Networks.

Rayleigh Quotient Graph Neural Networks for Graph-level Anomaly Detection

Detecting Anomalous Behaviours Based on System Logs: A Dynamic Graph Perspective