Core Concepts
Proposing a novel botnet detection model that fuses flow and topological features using a graph convolutional network (GCN) for improved performance.
Abstract
Introduction to botnets and their detection challenges.
Existing methods focusing on flow or topological features.
Proposal of a new model combining both feature types using GCN.
Pretraining strategy to address dataset imbalance for GCN training.
Experimental results showing superior performance over state-of-the-art models.
Real-world dataset validation and ablation experiments demonstrating the effectiveness of feature fusion.
Comparison of flow features effectiveness and optimal number of GCN layers for different architectures.
Classifier comparison highlighting Extra Tree as the best-performing model.
Stats
The accuracy of the proposed method is 98.85% under C2 architecture.
The recall rate achieved by the proposed method is 94.66% under P2P architecture.
Quotes
"Our model can effectively detect command-and-control (C2) and peer-to-peer (P2P) botnets."
"Our approach outperforms the current state-of-the-art botnet detection models."