A Supervised Machine Learning Framework for Early Detection of Man-in-the-middle Attacks

As the dependence on networked devices grows, protecting the Internet of Things (IoT) environment from Man-in-the-Middle (MITM) attacks has become essential. This study uses machine learning (ML) methods to strengthen the IoT landscape by providing a thorough investigation of MITM attack detection in IoT. By thoroughly examining traffic patterns, anomalous behaviours, and cryptographic protocols, the method seeks to identify aberrations that indicate possible compromises of machine translation mechanisms. supervised ML models are used to understand typical IoT device behaviours and quickly spot anomalies suggestive of MITM attacks. By constantly adapting to the changing nature of IoT surroundings, the suggested system not only identifies active attacks but also foresees and averts new dangers. Empirical studies and real-world simulations confirm the method's effectiveness and show how it may strengthen IoT networks' defences against complex cyberattacks. This paper expands the body of knowledge about strengthening IoT security by providing a proactive and flexible way to counteract the increasing threat of MITM attacks in the IoT space. The MITM attack detection dataset is utilised by the study to forecast MITM attacks that use different ML strategies, such as Random Forest (RF), Extreme Gradient Boosting (XGB), Decision Tree (DT), and Naive Bayes(NB). To find the best ML model, metrics like accuracy, sensitivity, precision, and F1-score performance have been applied. The accuracy rates of the DT, RF, and NB are 97.25%, 97.31%, and 95.55%, respectively, while the highest accuracy rate is 99.12% for the XGB classifier.