International Research Journal of Education and Technology

The increasing complexity and volume of network traffic have heightened the need for robust intrusion detection systems (IDS) to protect against evolving cyber threats. Traditional IDS approaches often struggle to effectively detect sophisticated attacks and relationships between entities within the network. This study addresses this gap by developing an intrusion detection system that leverages the power of Graph Neural Networks (GNN) to analyze the underlying relationships in network traffic and identify anomalies or malicious behavior. The primary objective of this project is to build an Intrusion Detection System using GNN that models network traffic data as graphs, where nodes represent devices or IPs and edges represent interactions or communication flows. The system aims to enhance intrusion detection by capturing complex, non-linear relationships among entities within the network. By utilizing datasets like CICIDS2017 or NSL-KDD, the GNN processes these graphs to extract high-level features for classifying traffic as normal or malicious. The system incorporates advanced GNN architectures, such as Graph Convolutional Networks (GCN) or Graph Attention Networks (GAT), to improve detection accuracy while minimizing false positives

.

Keywords: Hate speech detection, deep learning, convolutional neural network, natural language processing, content moderation, text classification