An Ensemble Classification Approach for Efficient Disease Prediction in IoMT

All over the world, chronic disease is one of the most common causes of death. To detect early disease, it is important to improve patient results. Early treatment can also significantly improve patient results. The main causes are drinking unclean water, an unhealthy diet, addiction, and many more to enroll the patient in chronic disease. The critical challenge is to analyze the data for early disease detection. In an IoMT setup, different sensors collect medical data and send it through different gateways to store it in the cloud. The main issue facing the healthcare industry is predicting the correct disease using data from medical devices in the form of sensors. The incomplete, noisy, inconsistent, error-prone data can compromise the performance of machine learning algorithms. Also, efficient prediction of machine learning models with greater accuracy is a prerequisite for such a setup so that medical practitioners can rely on the results, especially in cases where doctors are not available for physical examination. To address these limitations, we have proposed a methodology that pre-processes data and then applies ensemble techniques to predict the diseases. The comparison to the baselines indicates better performance in terms of performance metrics of Accuracy, Precision, and Recall.