Heart Attack Prediction System Using Machine Learning

Abstract

There is an unrestrained increase in the rate of heart strokes at puerile ages. So, there is a need to develop a system that can predict this disease at an early stage and hence preventive measures can be taken at the initial stage. It is infeasible for common people to take tests such as ECG on a regular basis and hence we require a system that is easily approachable and reliable both at the same time. This system automates the risk prediction. We propose a prototype that has an easy user interface and can predict the vulnerability to a heart disease with input of basic symptoms that include age, sex, chest pain, etc. We can use effective machine learning algorithms in the application, K-Nearest Neighbors algorithm is best suited for smaller datasets and decision tree gives accurate results for huge datasets.