Development of machine learning-based real-time medical internet of things framework

The novel coronavirus disease (COVID-19) has increased the need for new technologies such as the Internet of Medical Things (IoMT), Wireless Body Area Networks (WBANs) and cloud computing in the healthcare industry as well as in many areas. These technologies have also made it possible for billions of devices to connect to the internet, communicate with each other, and use new services that can be accessed from anywhere and anytime. In this study, a new IoMT framework consisting of WBANs is proposed. Analyzes are carried out using health big data from WBANs and fog and cloud computing technologies. Fog computing is used for fast and easy analysis, and Cloud computing is used for time-consuming and complex analyses. The proposed IoMT framework is presented with a heart disease and a diabetes prediction scenario. In the disease prediction process, predictions are made on Fog computing with fuzzy logic decision making. In cloud computing, Node-Red and Apache Kafka are used in real-time data flow, Apache Spark in real-time data analysis and MongoDB structures that can easily store big data are used. However, the Apache Spark machine learning library MLlib's Decision Tree (DT), Random Forest (RF), Gradient Boosting (GB), Logistic Regression (LR) and Support Vector Machine (SVM) machine learning classification algorithms are combined with real-time classification algorithms. Estimates are made by performing data analysis. In addition, the performance comparisons of the classification algorithms used are done. When the results are examined, it is seen that 64% accuracy performance for diabetes in Fog computing using fuzzy logic and DT, RF, GB, LR and SVM algorithms in cloud computing have accuracy as 78,77%, 77,40%, 84,93%, 80,14%, and 79,45% for diabetes, respectively. In addition to that offers 89,74%, 92,31%, 94,87%, 88,46%, and 93,60% accuracy for heart disease, respectively. In addition, the throughput and delay results of heterogeneous nodes with different priorities in the WBAN scenario created using the IEEE 802.15.6 standard and AODV routing protocol are also analyzed. However, Apache Spark real-time data processing performance, which is used in data analytics, is examined. All the results obtained proved the high efficiency and feasibility of the proposed approach.