IoTShield: Explainable AI-based Intrusion Detection System for IoT Networks

IoTShield is an interactive and explainable intrusion detection system designed for IoT settings employing machine learning and AI-driven explainability. It employs a Random Forest classifier trained on labeled synthetic network traffic to identify typical IoT attack patterns. It leverages model interpretability features such as SHAP (SHapley Additive Explanations) and Gemini LLM (Large Language Model) to provide users with natural-language-based simplified reasoning. This paper outlines IoTShield's methodology, architecture, system assessment, and application scope in real-world applications. IoTShield presents a transparent, scalar solution to IoT monitoring security in a real-time manner.