Use of Machine Learning and Random OverSampling in Stroke Prediction

A stroke is a medical condition where blood flow to the brain is interrupted. This is caused because of blood clots. Various factors, like hypertension, diabetes, heart disease, smoking, and body mass index level, result in a stroke. Almost every year, more than 15 million people have a stroke worldwide. Stroke is one of the major reasons of worldwide deaths. Hence prediction of a stroke becomes very crucial. This paper is focused on using various machine learning algorithms, namely, Support Vector Machine(SVM), Polynomial Kernel SVM, Sigmoid Kernel SVM, Gaussian Kernel SVM, Random Forest, and Logistic Regression. The paper also presents a comparison between different algorithms and accuracy scores. Random Forest gave the highest accuracy of 99.5%. Hence Random Forest is the best-suited model to predict Stroke.