Abstract—Industrial automation underwent changes through IoT technology advancements which created major security threats against widely used industrial communication protocols including Modicon Bus (Modbus). Investigating the deployment of advanced Machine Learning (ML) Models and Explainable AI (XAI) techniques represents this research's goal to enhance Intrusion Detection (ID) within IoT Modbus networks. A detailed assessment of IoT Modbus traffic included an evaluation between multiple models starting with Random Forest (RF) and including XGBoost, Gradient Boosting, AdaBoost, Logistic Regression and Support Vector Machines (SVM). RF displayed superior performance as an intrusion detection model by reaching a 98.32% accuracy level along with 98.41% precision and 98.32% recall metrics and 97.49% ROC AUC score. The strong precision rate of 93.91% together with ROC AUC value of 97.40% makes XGBoost a dependable model for use. Local Interpretable Model-agnostic Explanations (LIME) implemented increased the model’s transparency by revealing the critical predictive features through decision-making explanations. The study demonstrates ensemble models bring superior results because XGBoost and RF models showed better performance than alternative models for detecting malicious activities. The system benefits from LIME integration because it delivers both transparent features explanations and clear insights about what aspects affect model predictions thus generating more system trust. Through this research IoT security boundaries advance while the study offers operational solutions to protect industrial control systems against active cyber threats in practical environments.
 
Keywords—cybersecurity, Intrusion Detection (ID), Models and Explainable AI (XAI), Modbus protocol, data breaches


Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).