Detection of Cardiac Abnormalities and Heart Disease Using Machine Learning Techniques

The prediction of heart disease is a very challenging task in medical science, and it is essential to predict accurately for deciding future treatment. Almost 30 million peoples have died due to heart failure and different heart diseases worldwide. Internet of Things (IoT) and machine learning are the techniques that help to understand the heart's current condition. Various researchers have developed a system for predicting heart disease using several methodologies, but still, it remains a challenge to predict the accurate state of heart disease. The cardiac index and vascular age of the heart are the two significant vitals that indicate the precise condition of the heart. In this paper, we proposed heart disease prediction using IoT and machine learning techniques. Initially, we collected data from numerous sensors such as sunroom BP for heart rate, max30100 for blood oxygen saturation, EEG for PT and QR intervals, etc. The hybrid feature extraction and selection techniques and numerous machine learning algorithms have been used for strong training model building. With extensive experimental analysis, few machine learning (ML) and deep learning techniques have been evaluated with the existing implementation. The Recurrent Neural Network (RNN) obtains better detection and classification accuracy than conventional machine learning (ML) techniques such as SVM (Support Vector Machine), Naive Bayes (NB), Random Forest (RF), etc.