Analysis of ECG signal and Non-Invasive Blood Glucose Monitoring using Neural Networks

The metabolic condition known as diabetes, or Diabetes Mellitus (DM), is defined by consistently higher levels of blood glucose over an extended period of time. Diabetes management requires adequate blood glucose monitoring. Proper monitoring of blood glucose levels is essential for managing diabetes. Strong methods for tracking blood glucose levels have been the focus of research efforts. The amount of glucose in the blood can be estimated via NIR-spectroscopic measurement results. We first employed convolutional neural networks, yielding an r square value of 95 and an MAE of 0.16. Based on the numerical results, a non-intrusive blood sugar monitoring that incorporates near-infrared spectroscopy and an appropriate processing method emerges as a potential solution. The World Health Organization (WHO) conducted research revealing that the cardiovascular diseases (CVD) are the primary cause of death worldwide Cardiovascular-diseases (CVDs) account for almost 17.9 million deaths annually, or 31% of all fatalities globally. Experts believe that heart attacks and strokes account for 85 of these fatalities. One of the most common disorders that affect people is cardiac arrhythmia, which can result in life-threatening illnesses including stroke or cardiac arrest. Due to the asymptomatic nature of arrhythmia in many situations, one cannot determine whether they have it until they are checked by a professional. An ECG is a regularly used tool for monitoring a patient’s cardiac rhythm. The purpose of this work is to describe a technique that can be used to categorize immediately. We train and assess the proposed Convolutional Neural Networks (CNN) model on the MIT- BIH Database, yielding a 99% accuracy rate.