Simulation and modelling for anomaly detection in body sensor networks using machine learning classification techniques

Attacks and abnormalities in the sphere of Internet of Things healthcare have an influence on Body Area Sensor Networks (BASNs). Threats, malicious operations, and attacks in these infrastructures have developed in lockstep with the increased demand for body area sensor networks. Body Area Sensor Network anomalies have an impact on key decisions. We look at machine learning (ML) approaches to see if data sent to the server or collected by sink nodes in Body Area Sensor Networks is secure or not. In addition, we give an in-depth examination of a number of machine learning algorithms for the purpose of predicting a specific sort of anomaly using data from body sensors The anomaly detection technique is based on the Netsim simulator’s dataset. After proper dataset preprocessing, we show that XGBoost can achieve 97 percent test accuracy when compared to Decision Tree, Random Forest, and other classification algorithms.