OGIG: A Botnet Detection Model Integrating Graph Structural Modeling and Time Series Analysis

Botnets represent a severe cyber security threat, and their accurate and efficient detection is pivotal for safeguarding network security. This paper introduces a novel botnet detection model, termed OGIG, that synergistically integrates the strengths of OrderedGNN, Linformer, and Gated Recurrent Unit (GRU). Our proposed approach performs feature fusion by combining the structural features extracted by OrderedGNN and the temporal features extracted by GRU, which are subsequently fed into a classifier for botnet detection. Empirical evaluations on the ISCX-Bot-2014 dataset illustrate that our model achieves superior performance compared to other mainstream methods.