Effective Diabetes Mellitus Prediction Using a Hybrid Ensemble Machine Learning Model with Iot

Demand for healthcare Internet of things (IoT) devices that forecast and track illnesses has been driven by the rising worldwide prevalence of diabetes, which impacts more than 346 million people. A healthcare approach that does not involve invasive IoT monitoring and uses machine learning to forecast the likelihood of diabetes from key indicators like blood glucose levels is proposed in this study. The major objective is to create a program that assists diabetics in keeping their condition under control. Predicting the consequences of diabetes, the algorithm achieves 98.41 percent accuracy by using machine learning techniques, particularly bagging and boosting. For the purpose of training the model, 10, 221 users' data was culled from the PIMA database (PID), questionnaires, and IoT sensors. The IoT app provided a superior solution for predicting diabetes risk variables by integrating three conventional machine learning methods with a hybrid ensemble approach. This allowed it to outperform previous models.