Development of Machine Learning Model for Early Prediction of Cardiovascular Diseases among Type 2 Diabetic Patients using Electrocardiogram Signals

Abstract

Cardiovascular diseases (CVD) and Type 2 diabetes mellitus (T2DM) are two of the most widespread and severe public health challenges globally, with CVD accounting for 31% of all deaths and T2DM affecting over 537 million people. The coexistence of these chronic conditions significantly heightens the risk of adverse health outcomes, including myocardial infarction, stroke, and heart failure, and has become a leading cause of early morbidity and mortality newlinein patients with diabetes. In future the death rate due to CVDs may increase, these disorders are life-threatening and to avoid that early prediction of their symptoms need to be identified and necessitate immediate treatment has to be newlinegiven. newlineECG, a non-invasive and widely available diagnostic method, provides vital insights into the electrical activity of the heart, aiding in the detection of cardiac anomalies such as arrhythmias, ischemia, and hypertrophy. The state of the heart can be best understood through electrocardiogram analysis, which can be performed either manually or automatically to monitor it effectively. newlineManual diagnosis of electrocardiogram signals is challenging due to the presence of various morphologies within the signal. Thus, there has been interest in an automated electrocardiogram diagnosis system. The key components of any automated electrocardiogram classification system are feature extraction and classification techniques. newlineThe present research employs a range of machine learning approaches such as SVM, Decision Tree, KNN and Deep learning models such as LSTM, GRU to find ECG abnormalities among Type 2 Diabetic patients that leads to CVD. newlineConventional Machine Learning Methods is used to classify into normal and abnormal among CVD and Diabetic Patients. newlineBidirectional RNN model was proposed to classify and predict CVD and Diabetic dataset into early and abnormal Stage based on T- Peak. newlineHybrid RNN Model was proposed to classify and predict normal, early, abnormal stage of CVD and Diabetic based on T- Peak. newlineThe techniques adapt