Efficient early prediction and diagnosis of diseases using machine learning algorithms for IoMT data

As an essential part of the internet of things (IoT), the internet of medical things (IoMT) plays an essential role in the healthcare industry for the timely prediction of diagnosis of diseases to avoid chronic illness. Because of the massive information to be processed by the healthcare industry, some factors such as security, processing power, and accuracy of these information are of great importance for predicting the diagnosis of numerous diseases. To overcome these challenges, machine learning algorithms are used in the literature to increase the accuracy of patient's data. On the other hand, in this research work, patient data is collected from several IoMT devices such as ambulance, medical imaging, wearables, doctor reports, patient history, and labs. All data collected from several sources used in machine learning algorithms to categorize, cluster, and forecast for treatment and diagnoses. The provided results demonstrate that the random forest algorithm gives more than 93% accuracy, and the Hoeffding Tree algorithm gives more than 92% accuracy for patient heart data compared to other suggested algorithms in the literature. Several clustering algorithms are applied such as EM, k-means, density, filtered, and farthest clustering. K-means, filtering, and density algorithms give more reliable clustering results than others.