Improving Transparency In Intrusion Detection Systems Through LIME And SHAP-Based Explanation Of MLP Models

Abstract

Intrusion Detection Systems (IDS) are critical for safeguarding network security by
identifying malicious activities in real-time. While Multi-Layer Perceptron (MLP) neural
networks have demonstrated high accuracy in detecting intrusions, their complex decisionmaking
processes remain largely opaque, limiting their practical adoption. This study
explores the application of Explainable Artificial Intelligence (XAI) techniques—specifically
LIME (Local Interpretable Model-agnostic Explanations) and SHAP (SHapley Additive
exPlanations)—to enhance the interpretability of MLP-based IDS. By providing transparent
and interpretable explanations of individual intrusion predictions, these methods help security
analysts understand, trust, and effectively respond to alerts generated by the system. The
comparative analysis highlights the strengths and limitations of LIME and SHAP in terms of
explanation quality, computational efficiency, and applicability in real-world intrusion
detection scenarios. The integration of XAI with MLP models promises to bridge the gap
between high-performance detection and explainability, advancing the development of
trustworthy cybersecurity solutions