XAI Securenet: Explainable AI Techniques For Robust Intrusion Detection In IoT Infrastructures

The paper addresses the employment of machine learning models in Intrusion Detection Systems (IDS) in Cloud and iot settings, which produce high network traffic. Even though these models perform well, they tend to lack interpretability, which brings down the confidence level of security experts. The paper suggests employing shapley Additive explanations (SHAP), an explainable AI approach, to enhance the interpretability of ML models so that they are even more effective in identifying cyberattacks. This study explores the creation of a network intrusion detection system (NIDS) with a random forest classifier learned from the CIDDS001 dataset. The system produced a classification accuracy of 99.56runtime performance with predictive capability. The model was contrasted with baseline classifiers and evaluated in real-world settings with Wireshark. The research provides a solid, interpretable, and deployable NIDS solution for modern network security problems.