Explainable Artificial Intelligence (XAI)-Based Intrusion Detection and Classification in Network Traffic

Gorre Renuka Computer
Science and Engineering,Methodist College of Engineering and Technology, (OU Affiliated) Hyderabad,India
renukagorre80@gmail.com
Kudikilla Chandana
Computer Science and Engineering, Methodist College of Engineering andTechnology, (OU Affiliated) Hyderabad, India
chandanakudikilla@gmail.com
Anikha Tamkinath Computer
Science and Engineering,Methodist College of Engineering andTechnology, (OU Affiliated)Hyderabad, India
anikhatamkinath5@gmail.com
Mr. S. Chiranjeevi, Assistant Professor,
Methodist College of Engineering and Technology, (OU Affiliated) Hyderabad, India

ABSTRACTThe rapid growth of digital communication has increased the volume and complexity of network traffic, making modern computernetworks more vulnerable to various cyberattacks and data breaches. To safeguard these networks, Intrusion Detection Systems (IDS) are widely used to detect malicious activities within network traffic. Traditional IDS models rely on machine learning algorithms toclassify traffic as normal or malicious; however, they often act as black-box models, offering little to no insight into how predictions aremade. This lack of interpretability reduces user trust and limits their real-world applicability in critical security environments. Theproposed project introduces an Explainable Artificial Intelligence (XAI) approach that integrates XGBoost with SHAP (SHapleyAdditive exPlanations) to build a transparent and interpretable intrusion detection system. The model not only detects and classifiesintrusions accurately but also explains the contribution of each feature to the final prediction, enabling better understanding and trust. Thesystem is expected to achieve high detection accuracy while providing clear, visual explanations for each classification. This enhances decision-making for network administrators and contributes to a more secure and transparent cybersecurity framework. Keywords—Intrusion Detection System (IDS), ExplainableArtificial Intelligence (XAI), Machine Learning (ML), XGBoost,  SHAP (SHapley Additive exPlanations), Cybersecurity, Network Traffic Classification, Network Security Data Classification andModel Interpretability.