Symptom Wise Myocardial Infarction Prediction Using Machine Learning

Abstract

Cardiovascular diseases especially Myocardial Infarction (MI) (Heart Attack) is among the leading causes of death across the world. Early designation of this illness typically will increase the possibilities for productive treatment by that specialize in detecting symptomatic patients as early as potential. In this paper, we are trying to concentrate on symptom-wise heart attack prediction through machine learning approaches. Here, input data are collected from medical hospitals with a number of attributes like blood pressure, serum cholesterol, heart rate, BMI, troponin, diabetes, smoking etc. Finally, these features are modelled for prediction using some algorithms like Naive Bayes Classifier, Decision Tree, K-Nearest Neighbors, Linear Regression and Logistic Regression. Among all above mentioned algorithms, K- Nearest Neighbor gives the best prediction with 90% accuracy.