Identification of abnormal nodes in network communication based on graph theory algorithm
Gui Danping1，Fei Yang2
(1.School of General Education, Minnan Science and Technology University, Quanzhou 362300, China; 2.School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China)
Abstract： The traditional methods of identifying abnormal nodes in network communication, which rely on rules and signatures, or methods that only use partial graphical features, are limited when identifying key users. An anomaly node detection algorithm based on graph theory is proposed in this paper. Firstly local area network datasets collected offline are used to build a graph network; multiple graph features are analyzed to locate abnormal nodes in the network and analyze their potential abnormal behavior; secondly, experiments are conducted to test the detection effect on public network datasets. As a result of the final test results, it has proven to be efficient, convenient, and practical in locating abnormal nodes in network communication.
Key words : graph theory algorithm; abnormal detection; graph network generation; graph feature analysis
