Heart Attack Prediction using Machine Learning Approach

Abstract

For several decades, cardiovascular diseases have been one of the leading causes of deaths around the globe. Underlying health issues and lack in their timely detection highly contribute to the spike in mortality every year. It is unanimously agreed by healthcare providers, that early and accurate detection of diseases are essential to reduce the alarming mortality rate. With advancements in technology, research in artificial intelligence and machine learning algorithms, several studies have incorporated computer knowledge into healthcare industry. To cater this rising issue, this paper proposes an artificial intelligence-based model that aims to assist clinicians and cardiologists in predicting the possibility of heart attack in an individual. A machine learning framework is proposed utilizing a dataset that consists of 303 instances. The proposed method is analyzed using the ‘Heart attack Prediction’ dataset and the results obtained were robust. A comparative study was carried out which determined K-Nearest neighbor algorithm as the best approach; having an accuracy of 90.16% and recall of 87.09%. This study can be carried forward by utilizing machine learning models to predict/diagnose specific cardiac disorders, such as, right-heart disease(s) identification using Jugular Venous Waveform.